computer-notes

My notes about computer science, programming, and software
© 2008-2009 Jean-Marc Vanel

Last update:
This is my blog. Sometimes I say things in French there, and sometimes in english here.

2010-02-07

I'm annoyed by the Java bug in URL implementation : "URL class definition voilates RFC 1738 Section 3.10"

List of Projects That Use Protege

Gloze problems

The gloze.base parameter

The gloze.base parameter is "tolerant" when the "root resource" is <> : e.g. adding this is mandatory, but any URL can be provided, e.g. :

But when the "root resource" is not <> , the right gloze.base parameter is mandatory.

Apparently this is related to the Java bug in URL implementation : "URL class definition voilates RFC 1738 Section 3.10"

Jena URI ressources from the Jena API have file:/// , contrary to the JDK implementations of classes File, URL, URI.

Morewover, when reading an RDF file, the root resource must start with file:///, and gloze.base also. When reading an RDF file, Gloze tolerates a single slash in the resource URI for root in the file. So it's better to stick to triple slashes when authoring or generating an RDF for XML translation.

Problem with xml:lang in the XML Schema

With the very simple N3 file based on the XHTML Schema (see below), I got that stack :

Problem with the XHTML Schema and string content

Here a simple N3 file based on the XHTML Schema. This rather complex Schema is important in applications.

I got this rather long stack that is hard to understand. What is clear is that a Literal "bla" is cast as a Resource.

A bad warning message

[ WARN] - 2010-02-06 11:14:22,226 [AWT-EventQueue-0] com.hp.gloze
(BeanLoader.java:load:102)

This is due to a div element , but not in the input of Gloze, in the XML Schema, inside an <xsd:documentation> . In this particular case, that is not in the main XSD :

Gloze within EulerGUI

The XML rood node is an RDF blank node

First my code for "smart drop" was not good when the XML rood node is an RDF blank node ( _:sk0 ).

I did a correction in Gloze.rdf_to_xml() , but that was not enough. I got the "classical" error in method drop(Resource resource) :

2010-02-05

RDFa

2010-02-03

To read

phd_kaljurand.pdf ; Controlled English for Reasoning on the Semantic Web: decoi2009rewerse.pdf

Attempto Controlled English for Knowledge Representation : reasoningweb2008_fuchs.pdf

owled2008eu_kaljurand.pdf : ACE View — an ontology and rule editor based on Attempto Controlled English

http://www.faviki.com/ Faviki is a social bookmarking tool that lets you use Wikipedia concepts as tags.

2010-02-02

I found this called "Vim as Haskell IDE" : http://www.cs.kent.ac.uk/people/staff/cr3/toolbox/haskell/Vim/vim.html

Haskell

Haskell is rather mathematical, and mostly about functions, but it droped the essential of math functions syntax: the parenthesis!

<xerox> jmv - it's because the most important construct should use the lightest syntax: in this case function application, and the syntax is just juxaposition

Recent EulerGUI commits

2010-02-01

Recent EulerGUI commits

Refactor : extract the panel for buttons For an N3 Source to a new class ButtonsForN3Source

- for the second query created with orange ? button and after, it is put among the searches

- button for search should be different from the unique query button, without "activated" checkbox

A multi-valued property that is a LinkedHashMap: question on ##swing

Hi ! I have a multi-valued property that is a LinkedHashMap<String, Something> ,

is there some good practice, or some support classes , similar to java.beans.PropertyChangeSupport ?

2010-01-30

Specification of EulerGUI Server : parametrized queries

I wonder if there is a more of less standard way of wrapping a pre-cooked SPARQL query as a simple HTTP GET request.

It would have the query name, and possibly parameters that would be replaced in the query before running it. Here is discussion for parametrized (or parameterized ?) queries for SPARQL :

2010-01-27

Installed Thea

Running FuXi from repository: trouble

Links

2010-01-25

Readings: ACE View, Jena 2 Inference support

2010-01-22

Misc. links on OWL reasoners

(Re)Tried Open Mind / Common sense

I tried to teach him about carpel: it has parts: style, stigma, ovary (cf Flower Morphology ).

I wonder how to extract a part of the database, and also in which format it is available.

2010-01-21

Recent EulerGUI commits

When creating a new SPARQL query, with a non-existing file, the result of the query is put in a new /tmp file, but it was not available through the leftmost button.

2010-01-20

Announces EulerGUI 1.4.1

EulerGUI bug: creating a new SPARQL query

When creating a new SPARQL query, with a non-existing file, the result of the query is put in a new /tmp file, but it was not available through the leftmost button.

NOTE: I "simply" redo the layout, because the action for the leftmost button is not easily accessible. I could pre-create the /tmp file, or have an observer for changes on the N3 source. The RDF case is implemented differently :( .

2010-01-19

(re)trying Euler server

I simply started the class euler.Codd in the Euler distribution's directory /home/jmv/apps/Euler-2009-12-18 , with this argument :

2010-01-18

2010-01-13

Trying Joseki from sources

A server that wraps EulerGUI

More precisely, it should wrap a Drools Working Memory that can keep the user data for the duration of the session. The data persistence needs more thinking, as the first version will be a pure expert system.

Updated Fuxi

Subtleties of regular expressions

This article illustrates the subtleties of regular expressions, and their limits compared to compiler generators:

Drools N3 engine: factorial test

Move towards a normalization of strings and number when Triples are created (not in toString() and equals() ).

2010-01-11

RDF servers

It is not clear what is the difference between Joseki and TDB, as they both support the SPARQL Protocol and the SPARQL RDF Query language.

2010-01-10

2010-01-08

EulerGUI 1.4 Announce for English language followers

The main new feature in this release is plain XML import and export, based on XML Schema and a reasonable mapping between RDF and plain XML. The following XML Schema are hard coded (but you may ask for more): XHTML, SVG, MathML, XForms, XSLT, WSDL, DocBook, XBRL, Maven, Dublin Core in XML, UBL (from Oasis), KML (geography).

There is also a new N3 / RDF source: query from SPARQL services (using ARQ 2.8.2 from Jena project ) .

The OWL reasoner Pellet is integrated in a simple way, and shows OWL class tree for each source.

We now test the GUI at the functional level . It is happiness to see all these use cases that turn on their own quickly ! It now runs in about 3 minutes, but it should even be reduced by parallelizing the Drools compilations, which takes several seconds each. Jemmy is a test library for Swing and SWT. Jemmy has been used for years by the NetBeans team. The tests are written in Java, and JUnit tests added to the EulerGUI test set. These EulerGUI tests include the use cases of the Manual of Deductions Project : generation of applications from sources on Internet in languages N3, OWL, RDFS, UML, and eCore.

- First shot at integrating OWL reasoner Pellet: show class tree for all sources

- Drools N3 engine : built-ins: combine e:findall and math:memberCount; implemented greaterThan, notLessThan, notGreaterThan, notEqualTo, equalTo

- svn revision number and Class-Path are visible in Help / About ... as SCM-Revision:

EulerGUI 1.4 Announce on Protégé list

I said that the rule language is N3 logic . N3 logic is similar to SWRL in its semantics, but the syntax is similar to the body of SPARQL queries.

But I forgot to say that, contrary to SWRL, the predicates and classes do not need to be declared beforehand. There are 4 rule engines available, including one 100% Java based on the famous RETE engine Drools .

EulerGUI has a notion of project, including many kinds of sources that participate in the inference launch: N3, RDF, OWL/RDF, XMI/UML, plain XML, results of SPARQL queries.

The Euler Prolog engine can use rules in N3 logic for most of OWL concepts, see documents in owl-*.n3 in

http://eulersharp.svn.sourceforge.net/viewvc/eulersharp/trunk/2003/03swap/?sortby=file#dirlist

There is also a new N3 / RDF source: query from SPARQL services (using ARQ 2.8.2 from Jena project ) .

- First shot at integrating OWL reasoner Pellet: show class tree for all sources

- Drools N3 engine : combine e:findall and math:memberCount; implemented greaterThan, notLessThan, notGreaterThan, notEqualTo, equalTo

- svn revision number and Class-Path are visible in Help / About ... as SCM-Revision:

Drools N3 engine: built-ins

2010-01-06

Trying TDB Jena

Intelligent test generation

Dragomir D. liked my article "AI use for software development : the new paradigm". I'm reading his article quest to automated test generation .

Thanks to him, I knew NeON toolkit . I tried it , but the pre-release is unusable ... And the 1.2.3 release has no download for Linux 64 bits.

ordered lists in OWL

To model in OWL a route through several cities, I do not know what is good practice: use rdf:List , or rdf:Seq , or something else .

You are thinking at the rdf level of abstraction and not the owl level. If you use rdf:List or rdf:Seq then you will have an owl full specification. Reasoners generally will not work and it is very likely that the meaning of the ontology will not be the same for each reader.

Nick Drummond has given several presentations where he has described how to do what you want in OWL and exactly what you can say in OWL. A quick google search gives the powerpoint presentation [1] where he describes the approach. There is also an ontology that you can find here [2].

and particularize the class OWLList as a list of cities, by two owl:Restrictions :

What I regret is that the OWL List ontology is not a standard ontology , but it's not a big deal.

Further to the references that Timothy provided, there is a paper describing the list pattern:

Drummond, N., Rector, A.L., Stevens, R., Moulton, G., Horridge, M., Wang, H, Seidenberg, J. Putting OWL in Order: Patterns for Sequences in OWL, in 2nd OWL Experiences and Directions Workshop, Athens, GA

Links: Non-monotonic_logic, Closed_world_assumption, etc

http://granddictionaire.com/ propose pour call back, "procédure de rappel". Je dirais plutôt "rappel sur évènement"

2010-01-05

Attempto (ACE)

There is in fact some Prolog sources for SWRL in utils/drs_to_owlswrl.pl , but it is not accessible through the command line predicate ...

2010-01-04

Releasing EulerGUI 1.4

- First shot at integrating OWL reasoner Pellet: show class tree for all sources

- Drools N3 engine : combine e:findall and math:memberCount; implemented greaterThan, notLessThan, notGreaterThan, notEqualTo, equalTo

- svn revision number and Class-Path are visible in Help / About ... as SCM-Revision:

MVEL

I had never realized that MVEL used in Drools is in fact a separate language: http://mvel.codehaus.org/About. It will also be used in rule engine Take.

RIF (W3C Rule Interchange Format)

RIF is the W3C Rule Interchange Format. It's an XML language for expressing rules which computers can execute.

From this RIF Implementations list , I see that Euler implements RIF ! :) , using external RIF/XML->N3 translator . But where is this external translator ?

Attempto (ACE): tried some examples

I installed ACE without problem from the distribution. Then I can run commands e.g. (can also run on the the APE Web client without installing anything ) :

This last one coming from the RuleML tutorial is interesting because, calling ./ape.exe -solo paraphrase on it, one sees that the input is wrong. After some trials and errors I got this:

Note also that here "the discount" triggers this message on the APE Web client :

The definite noun phrase 'the discount' does not have an antecedent and thus is not interpreted as anaphoric reference, but as a new indefinite noun phrase.

Note also that here "applied" is interpreted as an adverb. This variant is propably better:

Tests with empathy and Google Talk

introductory material on the semantic web

I was looking for some solid links for resources and entry level to medium introductory material on the semantic web, RDF, and SPARQL. I like this intro by Marko A. Rodriguez: http://markorodriguez.com/Lectures_files/rdfrdfsowl.pdf

This was answered by Peter Neubauer, the guy of Neo Technology http://neo4j.org/

neo4j

2010-01-03

Checking EulerGUI dependencies

RuleML

For instance, datalog.xsd includes by schemaLocation 8 other modules. The hierarchy of models is detailed here:

RuleML and ACE

I was able to compile and launch the translator from ACE to RuleML. It calls the Web service for ACE (that parses the Controlled English), and prints the XML for RuleML in a JPanel. The class DRSParser (made with JavaCC from the grammar DRSParser.jj , described in thesis "TRANSLATOR: A TRANSlator from LAnguage TO Rules", by David Z. Hirtle, April 12, 2006 ) parses the DRS from ACE, and directly generates the XOM XML elements obeying the XML Schema for RuleML. The translation rules are described in the JavaCC grammar. The crucial lines are:

Attempto and N3

I'm trying to figure out how I can export an ACE rule base (in DRS, discourse representation structure ) as N3 logic. RuleML is just one possible intermediary. Directly using the ACE tools in SWI Prolog might be better. There is in fact some Prolog sources for SWRL in utils/ , but no command line predicate ...

I tried to re-use the DRS XML, by converting it with EulerGUI and Gloze. Alas the order of the 2 DRS tags inside Implication is not kept in the conversion, and there is no XSD available for DRS. I could follow the same path, but with the XML RuleML translation provided by the translator from ACE to RuleML .

An interesting article: Querying Ontologies: A Controlled English Interface for End-users

2009-12-24

Consolidating EulerGUI

2009-12-23

Gloze

2009-12-22

Gloze use cases

2009-12-21

Gloze

Maybe use this constructor to initialize Gloze with the important Schema list from EulerGUI :

2009-12-20

Gloze

2009-12-19

Gloze

2009-12-18

Recompiling Gloze 1.01-beta2

because vars was missing !!!!!!!! The debugger showed only the arguments, not the local variables.

On december 2nd, I tested Gloze with Ant (see Testing my modifications in Gloze ). Now I can't do ant compile anymore (but did I really run it ?) ! But ant test still runs !

Debugging Maven test cases with eclipse

First of all you need to get corresponding Maven projects into the Eclipse workspace.

To debug Maven standalone Maven CLI you can set the following environment variable:

Then start Maven as usual and attach remote debugger to Maven process on port 8000 using Eclipse's "Remote Java Application" configuration.

2009-12-17

EulerGUI

XML export with Gloze

Issues with "root" resource in an RDF graph

Gloze "drop" (RDF to XML conversion) now works on an N3 file, but a special one, starting with :

I discovered the role played by the gloze.base parameter . In the case of drop, this means the URI of the resource that will be converted to the XML root element. Indeed in a general RDF graph there is no such thing as a root.

I expected (naively) that Gloze was able to detect in a general RDF graph this "root" resource. But it does not, as it is not trivial.

So it has to be either given to Gloze, through the System Property "gloze.base" , or computed from the RDF graph and some information related to the XML Schema. The XML root element(s) for a given Schema is probably something that will have to be set in configuration (because an XML Schema does not specify a root element). Then to find the RDF statement that will be converted to the XML root element, one has to iterate on the statements, finding one that has the right predicate ( kml_:kml in the example above). So we could add in Gloze this method to set the RDF predicate of the "root" resource corresponding to the XML root:

The root RDF predicate would also be associated to a System Property "gloze.root" .

Arranged URI management in EulerGUI

Now at 1pm, it only works with eclipse; in fact there was 2 errors in file URI management:

In fact, given the message, and the fact that this particular Schema for KML does not declare the root element <kml>, it seems normal that it does not work.

The strange effect with Maven it certainly due to a mixture of static and non-static stuff in Gloze. Before me, Gloze was never tested a a reusable object.

After even more investigations, an EulerGUI test project with 2 sources, one N3 and one RDF, both using the KML vocabulary from the Schema, shows the same problem when exporting the second RDF source [on 21st could not reproduce that].

2009-12-16

We are a young team of researchers with backgrounds in information systems, conceptual modeling, economics, computer science, and related disciplines. In our research, we work on the theoretical and practical challenges of using the World Wide Web for facilitating the exchange of tangible and intangible goods between economic actors, and economic activity in general. Our core research interest is employing Semantic Web technology for E-Procurement and Business Process Management.

EulerGUI

XML export with Gloze

We need to reuse this instance of Gloze in class XMLExport, instead of the static method

2009-12-15

Fetching Jena sources from depots

EulerGUI

XML export with Gloze

As a side task I need to find a small XML Schema, preferably non trivial and useful. I try KML, a geographical Schema used in Google Earth, with this pair XML namespace + Schema location:

To test, I can click on XML export on a small KML document (eulergui/test/kml.xml) imported in EulerGUI. The question is: how can I find the pairs XML namespace + Schema location corresponding to an N3 /RDF documents ( typically coming from a rule engine ) ?

So we see that Jena RDF to N3 translator has created two N3 prefixes for KML. Note that there is a loss a information here, but I don't think that Gloze keeps the correspondance between the RDF predicate kml_:kml and the XML elements they come from. Note that the root <kml> element is not in the XML Schema for KML .

So we need to get access to the list of N3 prefixes, see if some of them match the XML namespaces in namespace2schema0 in XMLExport.java, and add the pair XML namespace + Schema location to the Gloze arguments.

In fact this is what XMLExport.selectSchemataFromProject() does :) . But for an N3 source coming from RDF, method N3Source.getKnownURIPrefixes() returns an empty set. In fact, even for an N3 source, this method returns an empty set when the Drools N3 engine has not been launched, i.d. when the N3 parser has not run yet. So I'm adding a "lazy" N3 parsing to method N3Source.getKnownURIs() .

2009-12-14

EulerGUI Drools N3 engine

Cardinality

2009-12-13

EulerGUI Drools N3 engine

Need to change that in Triple, because loading factsTriples.xml currently is broken .

Cardinality

In the Drools N3 engine , there was a confusion between the results and the rest; it was not serious as long as we did not count the triples. But this test, findall3.n3 , returns 3 instead of 2 :

2009-12-12

Installed Drools eclipse plugin

Note: on Ubuntu, the eclipse installation directory is /usr/share/eclipse/ , but I don't need that right now.

When asked to "Select a Drools Runtime", just click on "Configure Workspace settings ..." / Add / "Create a new Drools Runtime ..." , and then just give it a new directory.

EulerGUI Drools N3 engine: cardinality

2009-12-10

EulerGUI Drools N3 engine: cardinality

First commit for this combination of e:findall , math:memberCount , and math:greaterThan :

2009-12-09

EulerGUI Drools N3 engine: cardinality

Conmmited a big change to pave the way for cardinality rules in Drools N3 engine

2009-12-08

NIEM, the National Information Exchange Model, is a partnership of the U.S. Department of Justice and the Department of Homeland Security. It is designed to develop, disseminate and support enterprise-wide information exchange standards and processes that can enable jurisdictions to effectively share critical information in emergency situations, as well as support the day-to-day operations of agencies throughout the nation.

2009-12-04

2009-12-03

EulerGUI: SPARQL queries

I have a look in the ARQ ChangeLog to see what would be a replacement. I found this interesting page with the solution:

Enhancement in opening buggy projects

A Project with an N3 source throwing an exception at restore during prepare() does not prevent the project from being loaded; the N3 source is shown with a rose background and a tooltip showing the exception.

2009-12-02

EulerGUI : enhancements in tests

Bypass the Parser4J tests

The Parser4J N3 parser in EulerGUI is quite stable now. So, shorten the test time, I added this in 2 places in the Parser4J tests in EulerGUI:

Disactivate the windows in Jemmy tests

It works ! Note that I used to have close request sent by Jemmy to Firefox (bug of Jemmy, of my tests, of Firefox or of X-Windows ?). Of course this trick of running inside a second X server removes also this problem.

Testing my modifications in Gloze

http://eulergui.svn.sourceforge.net/viewvc/eulergui/trunk/eulergui/src/main/java/com/hp/gloze/Gloze.java

http://eulergui.svn.sourceforge.net/viewvc/eulergui/trunk/eulergui/src/main/java/com/hp/gloze/Gloze.java?view=log

2009-12-01

Testing Gloze in EulerGUI

There was a problem (a NPE) when adding one XML after another, both having the same XML Schema. Related to that, Gloze.main() was not made to run twice.

Question; can I reactivate the standard output when Jemmy has started waiting for stuff ?

2009-11-30

xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd

http://weblogs.java.net/blog/editor/archive/2009/11/24/stephen-colebourne-update-closures-jdk-7

2009-11-29

EulerGUI + Déductions : update Jemmy 2 test

Trying Jester, a test tester

Jester finds code that is not covered by tests. Jester makes some change to your code, runs your tests, and if the tests pass Jester displays a message saying what it changed.

The changes are very simple, like changing true to false, but the "succeded" changes point to missing tests, or in some cases to unnecessary code. Here is how I launched Jester. Because it re-builds after each change, it is quite slow, so I just ran one test and provided only one source class for changes

This means that only 21% of the mutations were caught by the tests. Actually in this package only one file is relevant : Instanciator.java , but a package directory could have been given instead, e.g. :

As an example, here is the result of Jester; a copy of the original source files as ".html" pages, with all the changes Jester made that did not cause the tests to fail shown in red, and the code it changed shown with a line through it : Instanciator.html

Joafip + Drools

Luc worked towards actual persistence. For now, it only runs in memory, but already uses the P* modified Joafip ollections.

2009-11-28

EulerGUI + Déductions : update Jemmy 2 test

It is nice for demos to keep the GUI active for interactions after scenario is played.

Add to one test the EulerGUI features: display class hierarchy with Pellet, plain XML as N3 source, SPARQL query, export as Drools, export as XML;

and also the generated aplication features: click on object properties, round trip with use N3 file, ...

2009-11-27

EulerGUI + Déductions : update Jemmy 2 test

I realize that the Jemmy API is like a scripting API for any Swing application. I should open the Euler GUI Manual and write the script for all the features.

2009-11-26

EulerGUI + Déductions : update Jemmy 2 test

I update Jemmy 2 test to include the main Déductions use cases: generate application from URL : N3 , OWL , RDFS , UML , and eCore documents.

While doing that, I met the annoying problem that the RDFS URL is long to download (in a separate thread), but the Drools engine does not wait.

I also set the right entry in Maven depot for Grappa ( this will accelerate every build ) :

On phone with Olivier Moïse

2009-11-25

EulerGUI + Déductions : added a Jemmy 2 test

Apparently, the presence of the other windows from previous tests are a nuisance.

2009-11-24

If you add; &fmt=18 in the end of each video on youtube, you will get the high quality version of the video. Works on every movie.

Looking for java gui testing frameworks

JToolbar problem with uispec4j

but then I loose the facilities offered by uispec4j to select further inside the JToolbar .

2009-11-23

My PS1 for zsh

2009-11-22

IDE's and tools

I tested (not latest) NetBeans available via Ubuntu. The import of my eclipse projects failed, because I have 2 eclipse projects on the Subversion depot ( trunk and a branch).

Questions about Abbot

Working on Joafip + Drools

2009-11-21

Working on Joafip

To upload on remove repository, just change the url pamameter, and add repositoryId:

The repository Id refers to the following content in <HOME_DIRECTORY>/.m2/settings.xml :

Putting Abbot on Maven

Download sources and build

Work on local repository without pom

I don't want to pollute my remote Maven repository for these trials. So I first did this to upload on the local Maven repository the jar from the Abbot Sourceforge project:

Work on local repository with pom.xml

Then I go further and use the pom.xml for Abbot I made from this discussion on the Abbot list :

In fact there are 2 artifacts: the core Abbot, and the recorder GUI, Costello, depending on Abbot. For EulerGUI we will need both.

Note that the groupId, artifactId, version and packaging informations are automatically retrieved from the given pom.

Using Abbot in my EulerGUI project managed by Maven

Then I can test the local Maven installation for Abbot, adding this in pom.xml :

Of course, I have another problem. My tests were recorded with Costello, and depend on some absolute positions.

2009-11-20

EulerGUI TODO

Rules for form generation

There should be a way to infer the action for events, based on a semantics added to the fields.

2009-11-18

Rules for form generation

Migrated to Pellet 2.0 ; show leverage on: incremental classification, fine-grained inference extraction (?), SWRL rules, SPARQL queries

2009-11-17

My Philips webcam works with cheese!

Rules for form generation

2009-11-16

Rules for form generation

Work on creation button: in the new window there was a extra form for the main class;

also for focus lost events there is often the creation of a new object and form.

2009-11-14

bddbddb stands for BDD-Based Deductive DataBase. It is an implementation of Datalog, a declarative programming language similar to Prolog for talking about relations. What makes bddbddb unique is that it represents the relations using binary decision diagrams (BDDs). BDDs are a data structure that can efficiently represent large relations and provide efficient set operations.

I probably made the wrong move when upgrading to Jena 2.6.2,because it is incompatible with Pellet rc7 . So I have to wait for the next release of Pellet, says Thorsten Möller.

Rules for form generation

2009-11-13

Webcam on Packard Bell

Rules for form generation

Also tested test_person_import.n3p , foaf_import.n3p, cv.n3p (no fields at all ), uml.n3p.

2009-11-12

Aleph & Inductive Logic Programming

2009-11-05

Explanations about rules for form generation

Tidied the pom.xml after upgrading to Jena 2.6.2 : remove slf4j-api and slf4j-log4j12 .

Writing Business Rules Engines in Mercury: Ian MacLarty Functional Programming Union, University of Melbourne, 24 July 2009. Available here (156K).

2009-10-30

Keyboard problems on ubuntu

I finally discovered that my minus and vertical bar are activable with Windows key ! Equal is with = and control key on terminal and eclipse.

Prepare EulerGUI 1.4

Grunt at subversive

For the latter, I used command line svn ( version 1.5.4 (r33841)) to download, and made an eclipse project from that.

Ubuntu 9.10

2009-10-29

Solved strange random errors in Maven tests for EulerGUI

(done this manually with gvim -d and kompare, as subversive eclipse is sooooo slooow)

In fact there was a modification in Instanciator introduced recently, that makes an error in the test. This modification removes the error messages in the JavaScript console.

Pause in Drools-Joafip-EulerGUI integration

He does not a have on Windows with JDK 1.6_18 the error I have on Linux with JDK 1.6_16:

2009-10-25

Hyperlinks: ubiquity-xforms, Nepomuk, transmodel.org, ...

2009-10-24

Create a branch in EulerGUI for joafip + Drools

This is the first branch in EulerGUI, so I must create the directory branches first:

Alas today (saturday!) the Maven site http://jena.hpl.hp.com/repo/ for Jena is awfully slow.

2009-10-23

Adapt a persistent storage to Drools ?

To integrate into Drools we would need a DAO design pattern, or the possibility to plug a specific implementation of ObjectStore (or SingleThreadedObjectStore).

2009-10-22

EulerGUI : on semanticweb.org; group in French

Discussions en Français pour les développeurs et architectes intéressés par les évolutions, les choix et la feuille de route des projets EulerGUI (http://eulergui.sourceforge.net/) et Déductions ( http://deductions.sourceforge.net/ ):

EulerGUI

TODO : does not work with JavaScript instanciator: end-of-lines in literal strings in N3

2009-10-21

Trying Virtuoso RDF store, Open Source version

I took the 6.0.0 version, that has no release notes :( . The licence is GPL v2. Then, according to the README, I missed these packages :

Since I have not created any other user, I upload an OWL file as dba with an HTTP PUT :

The trouble is that this OWL file was uploaded as a binary, as it was sent with Content-Type application/octet-stream. You can see this by clicking on WebDAV Browser.

Complete Kubuntu to Ubuntu migration

After this site, here are the packages to remove for a Complete Kubuntu to Ubuntu migration

sudo apt-get remove akregator amarok amarok-common apport-qt ark cdrdao dolphin dontzap dragonplayer exiv2 foomatic-db-gutenprint gdebi-kde gnupg-agent gtk2-engines-qtcurve gwenview hpijs-ppds ijsgutenprint install-package jockey-kde k3b k3b-data kaddressbook kamera kate kde-icons-oxygen kde-printer-applet kde-style-qtcurve kde-window-manager kde-zeroconf kdebase-bin kdebase-data kdebase-plasma kdebase-runtime kdebase-runtime-bin-kde4 kdebase-runtime-data kdebase-runtime-data-common kdebase-workspace-bin kdebase-workspace-data kdebase-workspace-libs4+5 kdebluetooth kdegraphics-strigi-plugins kdelibs-bin kdelibs-data kdelibs4c2a kdelibs5 kdelibs5-data kdemultimedia-kio-plugins kdepasswd kdepim-kresources kdepim-strigi-plugins kdepim-wizards kdepimlibs-data kdepimlibs5 kdeplasma-addons kdeplasma-addons-data kdesudo kdm kfind khelpcenter4 klipper kmag kmail kmix kmousetool knotes konqueror konqueror-nsplugins konqueror-plugin-searchbar konsole kontact kopete korganizer kpackagekit krdc krfb ksnapshot ksysguard ksysguardd ksystemlog ktimetracker ktorrent ktorrent-data kubuntu-artwork-usplash kubuntu-default-settings kubuntu-desktop kubuntu-docs kubuntu-konqueror-shortcuts kuser kvkbd kwalletmanager language-selector-qt libakonadiprivate1 libao2 libaudio2 libavahi-qt3-1 libboost-program-options1.35.0 libclucene0ldbl libdbus-qt-1-1c2 libeet1 libexiv2-5 libflac++6 libgeoip1 libk3b3 libk3b3-extracodecs libkcddb4 libkdecorations4 libkdepim4 libkexiv2-7 libkholidays4 libkipi6 libkleo4 libkonq5 libkonq5-templates libkpgp4 libksieve4 libkwineffects1 libloudmouth1-0 liblua50 liblualib50 libmad0 libmimelib4 libmodplug0c2 libmpcdec3 libmsn0.1 libmysqlclient15off libokularcore1 libpackagekit-glib11 libpackagekit-qt11 libphonon4 libplasma3 libpoppler-qt4-3 libpq5 libqca2 libqca2-plugin-ossl libqedje0 libqimageblitz4 libqt3-mt libqt4-assistant libqt4-core libqt4-dbus libqt4-designer libqt4-help libqt4-network libqt4-opengl libqt4-qt3support libqt4-script libqt4-sql libqt4-sql-mysql libqt4-sql-sqlite libqt4-svg libqt4-test libqt4-webkit libqt4-xml libqt4-xmlpatterns libqtcore4 libqtgui4 libqzion0 libraptor1 librasqal1 librdf0 libsearchclient0 libsoprano4 libstreamanalyzer0 libstreams0 libstrigihtmlgui0 libstrigiqtdbusclient0 libvncserver0 libxcb-shape0 libxcb-shm0 libxcb-xv0 libxine1 libxine1-bin libxine1-console libxine1-misc-plugins libxine1-x libzip1 mysql-common okular okular-extra-backends openoffice.org-kde openoffice.org-style-crystal oxygen-cursor-theme packagekit packagekit-backend-apt phonon phonon-backend-xine pinentry-gtk2 pinentry-qt4 plasma-widget-network-manager plasma-widget-quickaccess python-dev python-kde4 python-packagekit python-plasma python-qt4 python-qt4-common python-qt4-dbus python-sip4 python2.6-dev qt4-qtconfig quassel quassel-data raptor-utils redland-utils software-properties-kde soprano-daemon speedcrunch strigi-client strigi-daemon system-config-printer-kde systemsettings ttf-dejavu ttf-dejavu-extra update-manager-kde update-notifier-kde && sudo apt-get install ubuntu-desktop

2009-10-20

Biological Collections software

Open Source; in Python; easily instable on Ubuntu. Alas on a fresh database several actions lead to an internal error. The families are builtin, but completion does not always occur. Could not make the report work. Note: I'd love to make a similar applicatoin with an RDF database and my GUI generation techniques.

Generating XForms with EulerGUI

For XForms (and XHTML?) output, I think about using the other, more abstract, layer in my rules stack, "gui_generic.n3" (see also Supporting ontologies).

Capturing desktop session for demos

Looking for tools for doing JavaScript animations

Troubles with webcam Philips SPC210C on Ubuntu x86_64

2009-10-19

EulerGUI Jena output stack trace

Generating XForms with EulerGUI

Here is the kind of XML that must be generated in the end of the tool chain (after the RDF to XML conversion by Gloze) :

Since XML elements and attributes correspond through Gloze to RDF / N3 predicates, the kind of N3 that must be generated is (with all blank nodes explicited):

Currently, the corresponding fragment of the generic GUI (actually very Java-oriented) that is produced before applying the N3 rules for Java translation :

The correspondance is indicated above with the color code: the panel becomes the subject of xf:textarea.

So, in plain english, I must for a panel having a gengui:TextField such as _:sk_JTEXT_FIELD_36, add a triple with the xf:textarea predicate:

Of course the big trouble is the xf:ref , bacause the XML data model of XForms is not RDF.

Links: TAP, etc

http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B758F-4B4X4HF-7&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1051473659&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=1f3210ebc45ea17df5f5d3dcf43987ad

Solving problems with 2 screens on Ubuntu

2009-10-16

Unix_way_of_sex = unzip; strip; touch; finger; mount; fsck; more; yes; gasp; umount; sleep

EulerGUI features

D'n'D of a Project URL, the N3 source buttons were not there, due to method D'n'D of a Project URL, due to method makeButtonsPanel()

2009-10-15

Déductions & XForms

These lines in deductions/n3_new/generic_gui_projection-rules.n3 are a problem, because they are much too Java-oriented:

2009-10-14

The tests for N3 ParserJ4 from rdflib

EulerGUI feature: read plain XML from URL

needed to adapt Gloze to process input URL's instead of Files : created GlozeURL as a replacement of Gloze

Although it's not implied by your License, it's a pleasure to share my modifications.

First I needed to adapt Gloze to process input URL's instead of Files, so I created GlozeURL as a replacement of Gloze :

http://eulergui.svn.sourceforge.net/viewvc/eulergui/trunk/eulergui/src/main/java/com/hp/gloze/GlozeURL.java?view=markup&pathrev=894

The modifications are rather systematic, replacing File by URL in the right places. So it should be really a replacement of original Gloze, and work whenever original Gloze works.

I made class GlozeURL extend Gloze, to avoid compilation errors and propagation of modifications elsewhere. But if you agree to use GlozeURL as the new Gloze, you should just remove the inheritance and rename GlozeURL to Gloze, and it should be OK. While I was here, I added generic arguments in a few places to avoid yellow markers in eclipse, and added curly braces for if blocks, that most code checkers object to.

Second, I've made a "short" list of important XML Schema that are automatically added to Gloze argument, depending on XML namespaces present in input file.

http://eulergui.svn.sourceforge.net/viewvc/eulergui/trunk/eulergui/src/main/java/eulergui/export/XMLExport.java

The automatic addition to Gloze argument, depending on XML namespaces present in input file, occurs here in class N3SourceFromXML_Gloze :

http://eulergui.svn.sourceforge.net/viewvc/eulergui/trunk/eulergui/src/main/java/eulergui/inputs/N3SourceFromXML_Gloze.java

There is a small SAX ContentHandler that just records the XML namespaces present in input file.

2009-10-13

EulerGUI features

Detect RDF, OWL and plain XML

TODO : to be able to download a plain XML from an URL, need to detect such root elements in SourceFactory and DropURLTranferHandler:

XML import

extracted XML Namespaces from plain XML source and transmit them with Schema location to Gloze.

2009-10-12

Déductions GUI rules: generate XForms

I don't know yet how easy it will be... Anyway it is a good example of generating XML.

2009-10-11

Isaac project: IsaacOS : object prototype based operating system, Lisaac : object prototype based language

EulerGUI new feature: XML Export

I need to expose the N3 prefixes to the XMLExport class. Indeed Gloze iterates on the given XML Schema for the RDF to plain XML translation, so it's not possible to add all the known XML Schema to the context. A partial list of of the most famous and useful XML Schema: XSD itself, XHTML, XForms, XSLT, SVG.

EulerGUI is a hybrid creature

After more reflection, EulerGUI (http://eulergui.sourceforge.net/) is a hybrid creature.

I suggested it in the category "rules engines" because it wraps Drools as a rules engine for the N3 logic language (http://www.w3.org/DesignIssues/N3Logic), and because it integrates 4 rule engines.

EulerGUI also has importers for models and data in UML, eCore, plain XML, XML Schema. So it could well be in category UML.

2009-10-09

EulerGUI new feature: XML Export

2009-10-08

Xorion : a CMS with XRX architecture

Xorion will be an open source, XML-database based content/community management system that makes extensive use of XQuery and related technologies. The goal is to create a Drupal-like community project that can be used with a RESTful Services architecture to host rich XML documents as well as blogs and other components.

Communication

2009-10-07

2009-10-06

Trying dbpedia

2009-10-05

Trying Sesame 2.3-pr1 SPARQL server

Trying Joseki SPARQL server

I have to see how to add documents in the database, and also try TDB (TDB is a component of Jena. It provides for large scale storage and query of RDF datasets using a pure Java engine. TDB supports SPARQL.).

EulerGUI new feature: calling SPARQL end point

The feature is implemented and documented, but I still have to find some SPARQL public endpoint with responses for CONSTRUCT queries.

- write new paragraphs: Show N3 parse tree, Show original source of N3 document,Launch OWL reasoner, SPARQL queries as N3 sources

Exploring SPARQL public endpoints

To start with nothing but a SPARQL endpoint and to dive in and learn about the data, the basic mechanism by which we can do this is by writing queries that use variables to find all of the predicates and all of the types that exist in a dataset, and then to pick out interesting predicates and types and use open-ended queries to explore the structure of the data. Dean Allemang has written a blog post on this exact subject, so I'll gladly reference his writing on using SPARQL to explore an unknown dataset.

2009-10-05

2009-10-01

EulerGUI new feature: calling SPARQL end point

Maven + Pellet

2009-09-30

EulerGUI new features

Integrating Pellet

TODO

EulerGUI doc. date

2009-09-29

Trying DL queries in Protégé 4

Trying Pellet SPARQL-DL queries with Pellet

EulerGUI : change the description on main page

Provides a GUI for the eulersharp.sourceforge.net Euler/SEM reasoning engine. The sources can be N3 (Notation 3), RDF, OWL, UML or eCore, files or URL's. It also wraps Drools as an N3 rules engine. It is becoming a lightweight IDE for Artificial Intelli

A lightweight IDE for Artificial Intelligence. Started as GUI for the Euler reasoning engine. The sources can be N3, RDF, OWL, UM, eCore, plain XML or XSD, files or URL's. Wraps Drools (or CWM, FuXi) as N3 rules engines. Model based app. generation.

2009-09-28

XML & XSD import : update

Now both ReDeFer and Gloze are integrated in EulerGUI (in Subversion), so it's more simple to compare. The output is quite similar in both. I corrected <xsd:all> in ReDeFer. I did nothing in Gloze, except a bit of defensive programming in two Jena classes. The problems I'm aware of are:

EulerGUI: specifications

Open a form "Déductions style"

This is quite easy to create a EulerGUI Project, add to it the current N3 source, and the swing-rules2.n3p project, then run that project (need to set the edited File to the current N3 source). The point needing more reflexion is "guess where are the RDFS or OWL sources that define used properties and prefixes".

Export whole project as XML

Export whole project as application

An equivalent Web application

2009-09-27

EulerGUI: Reading plain XML files or XML Schema

Mail to Roberto G.

2009-09-25

Links: Java sources

Maven ecosystem

Saxon B (now at version 9) is not up-to-date on Maven central, and not anymore on ibiblio :

cimtool.org

add Maven config. in ReDeFer

add Maven config. in ReDeFer (not tested; for Maven build to work, it will need some directory changes in sources)

Crash with Jena 2.6.0 and HSQLDB

This happens the second time the application is started, as if Jena was not aware of the real database being opened .

This means that I'm forced to debug Jena just because the newest stuff is not in Maven!

2009-09-24

2009-09-23

Déductions: plain XML as RDF

Testing Redefer plain XML to RDF conversion

XML2RDF is not based on a XSLT. It is implemented using Java and the Jena library. The code is available from

Problem with attribute namespace

According the XML namespaces Rec., the attribute ref in the input XML is in no namespace (not in default namespace in scope !). I propose to translate it with an RDF prefix correcponding to the original base URI.

2009-09-22

Missing feature in Gloze: enumerations

In the Protégé OWL 2 tutorial I saw the way to create Enumerated Classes in Protégé 4 (see parag. 7.3 Enumerated Classes).

2009-09-21

triplr

The content at the original URIs must be readable as one of the syntaxes Raptor knows - RDF ones, RSS, Atom, Microformats that have a GRDDL profile and so on. The full list is given in How... can I use it?

EulerGUI instantiation: correction

For instanciating a Method call with a list as object, a resource was wrongly taken in account as a literal string.

Déductions GUI rules: migration after package change in EulerGUI

Before doing a new release of EulerGUI, I put the Déductions GUI rules in a new directory:

2009-09-20

EulerGUI : refactoring : moving Project and N3Source in a separate package

Java instantiation in EulerGUI

A triple involving a Java object in first position (subject) will be interpreted as a JavaBean property, e.g. :

Precedently only predicates with the Java prefix were interpreted as Java properties.

Also method calls are now detected, no matter what the N3 prefix is, as long as it ends with # :

Trying rhizomik / Redefer project

2009-09-19

From OWL to Plone : I found a working path to transform an OWL ontology into a working Plone content-type. Here is my recipe

Déductions GUI rules: migration to generic GUI platform

I created a new variant of the basic Person test, person_import2.n3p, that imports a new variant of the basic Déductions GUI rules, swing-rules3.n3p.

For now only the classes are modified, not the java-specific predicates. The generic GUI platform objects are tagged as Java objects according to the chosen platform, e.g. :

2009-09-18

Déductions: plain XML as RDF

Tonight I realized the big potential of having a 2 directional plain XML <--> RDF correspondance. On input, a plain XML can be added to the RDF store. For instance, it can come from an XForms form. On output, it can solve the problem of generating plain XML out of an RDF object graph. And this last feature is needed for many use cases, as plain XML is ubiquitous today (feed an instance for an XForms, XBRL report). XBRL in Cover pages .

Trying to debug the output of Gloze

Trying rhizomik / Redefer project

Examine XForms for integration with Déduction GUI rules

"required" and "relevant" properties are interesting, as ell as "constraint" and "calculate". All are attributes of the <bind> element.

The contruct <repeat nodeset="/some/xpath"> takes care of multi-valued domain properties.

2009-09-17

Déductions GUI rules: migration to generic platform

Links: Dbin

Tried Gloze on a large XSD

Sent a mail to Gloze author Steve Battle, and also a question on Protégé list:

It's worse with Protégé 4 (build 113). I just get a stack trace. Is there a way to get a line number in the document without using a debugger ?

2009-09-16

Installing Sonar

Sonar enables you to collect, analyze and report metrics on source code. Sonar not only offers consolidated reporting on and across projects throughout time, but it becomes the central place to manage code quality.

2009-09-15

Exploring Jena

Exploring Jena code base

TDB is a high performance, pure-Java, non-SQL storage subsystem for Jena that scales to billions of triples.

Trying jena eyeball

I downloaded Eyeball from CVS. Then I added this small script run_main_class.sh :

Amaya : features wished

Fuxi

Fuxi and N3 syntax

Fuxi has moved

Chimezie has decided to separate the expert system (FuXi) into its own Google Code project and will decide soon what to do with python-dlp (the idea, anyways): http://code.google.com/p/fuxi/

I updated FuXi integration in EulerGUI, now that FuXi has the --filter feature. I was not without pain; first at some time the main script changed from Fuxi to FuXi; then rdflib 2.4.0 provided by Ubuntu was too old. I tried to summarize the install in the EulerGUI Manual .

2009-09-14

Trying SGML/XML with Yap

sourceforge download stats KO

There is currently an issue preventing download stats from being displayed properly. All the download data is being recorded, it simply isn't being displayed. This is a known issue, and we're working on a fix for it.

Trouble with Philips Webcam SPC210NC

Links:

http://optimix.sourceforge.net/ Optimix lets the user specify graph transformations in a declarative language, based on relational graph rewriting.

http://wwwcs.uni-paderborn.de/cs/varlet/tour.html The purpose of this guided tour is to demonstrate some functionality of the Varlet database reengineering environment.

2009-09-04

My program for "vacancies"

Announces for EulerGUI 1.2.1

Announce in the Protégé list

Déductions is a Java Swing application generator from OWL models and N3 logic rules.

This is a framework (Open Source, compatible but not Protégé-based) allowing to mix ontologies and a RETE forward chaining engine (Drools). This a quite similar to the SWRL plugin of eclipse, except that the rule language is N3 logic . N3 logic is similar to SWRL in its semantics, but the syntax is similar to the body of SPARQL queries. N3 is a good candidate for a universal data + models + rules + queries textual language. The RDF subset of N3 is perfectly isomorphic to RDF, but as human readable as can be. And it is readable by Protégé, Jena, and others tools. OWL, being expressed as RDF, is thus also expressible in N3. Here is an example:

For a developper, and even some domains experts, it can be quicker to enter initial models and facts in N3 with an editor having syntax coloring. Then Protégé can be used to visualize, navigate, update, and Description Logics reasonning.

EulerGUI is a lightweight IDE for developping and testing projects composed of N3, OWL, RDF(S) and UML ontologies and databases, and rules in N3 logic.

The rules are in N3 logic; so the same rule language is used for infrastructure (form generation) and for business rules. There are four N3 rules engines: in Prolog, Python (Fuxi, CWM) , Java (Drools - RETE) .

In 2 clicks EullerGUI allows to generate and run Java Swing GUI's from OWL models and rules from the Deductions project.

Announce for the Drools list

In short, EulerGUI is an IDE centered around a Semantic Web language called N3. N3 is a good candidate for a universal data + models + rules + queries textual language. The RDF subset of N3 is perfectly isomorphic to RDF, but as human readable as can be. And it is readable by Protégé, Jena, and others tools. OWL, being expressed as RDF, is thus also expressible in N3. Here is an example for an OWL model :

As you can see, a negation of the consequent (RHS) is added in the antecedent, to avoid extra responses.

The rule base, the models, and the data are exportable for using with Drools independently. So, for the Drools community, EulerGUI allows to import models and data from N3, RDF, OWL, UML, eCore documents. Also for some people, writing N3 rules can be easier than Drools language.

As you see above, the translation makes use of a particuliar form of Drools language, which makes use of a single Java class, Triple, that models RDF statements. This is quite far from the intended Drools usage, but it shows the power and flexibility of Drools. When embedded into generated Java Swing applications, a Drools engine is able to generate a form in a few milliseconds from the class model! Also this design, based on a single Java class, allows to add and remove business classes at runtime.

Déductions is a Java Swing application generator from OWL models and N3 logic rules.

Announce for the PlanetMDE list

Announce for the news AI list

Announce for news:fr.comp.lang.java

Announce for LinuxFr.org

2009-09-03

Déductions generated applications : test and debug

update convert-UML2_metamodel_to_OWL.n3p to download directly the UML.ecore from the Web

TODO: it is not currently possible in EulerGUI to set an URL for an eCore document, so the convert-UML2_metamodel_to_OWL.n3p project was edited by hand.

Test cases for Déduction from Web

2009-09-02

Project Sequence : postProcessingURL was not taken in account when it's a relative URL.

How to find any eclipse source file

Converting UML 2 metamodel into OWL

After correcting a slow algorithm and adding N3 escape """ to strings if necessary, I was able to use EulerGUI to convert the UML 2 metamodel into OWL. This is just a by-product of being able to express eclipse eCore models in N3. The funny thing is that the UML-OMG guys want to translate the Semantic Web things into their stuff (see OMD), and the Semantic Web guys (like me) do the converse.

XMI --> N3 representation of the XMI --> conversion through N3 rules --> OWL N3 representation

This simple eCore to OWL translation processes: Class, eSuperTypes, name, eStructuralFeatures. These eCore features are left aside for now: lowerBound, upperBound, eOpposite, containment="true", derived="true".

The resulting OWL converted from the UML 2.2 metamodel allows one to navigate into this big metamodel in Protégé, to write rules relying on UML metaclasses and predicates, and to enter instances (actual UML user models) in Protégé.

To make a complete translation, one should also translate the OCL constraints and operations, that could be converted into SWRL or N3 logic. This is quite a bit a work, but the OCL parser from eclipse is usable.

There might still be some imperfections from the point of view of UML or OWL experts.

I put the N3 file for the OWL conversion of the UML 2.2 metamodel (suitable for Protégé) here:

On the web the UML2 metamodel in the eclipse eCore XMI dialect, that was used for the conversion, is here :

2009-09-01

2009-08-31

Instead of monitoring several forums and threads , it is simpler to monitor Jos De Roo's profile:

Translating Java class into OWL

Above I follow paragraph "7.2 Abstract Syntax" in "OMG Unified Modeling Language (OMG UML),

Superstructure, V2.1.2", except that I use XSD datatypes and rdfs:label as usual in OWL.

Convert UML2 metamodel into OWL

First I took the eCore version of the UML 2 metamodel from my eclipse 3.5 installation:

2009-08-30

2009-08-28

OWL 2 with form generator: TODO

Google Summer of Code 2009

Prolog

Prolog design patterns

2009-08-26

Pyke, Logic Programming in Python

Pyke introduces a form of Logic Programming (inspired by Prolog) to the Python community by providing a knowledge-based inference engine (expert system) written in 100% Python.

Spell check in different language in Amaya

To change the language, apply a lang attribute to the text with the Attributes tool, and restart the spell-check procedure.

2009-08-25

Links: fresnel, etc

Releasing EulerGUI with Maven

Because the pom.xml was not in a good state, I had to run release:prepare this way :

and commits the copy of the sources in the tags/ directory in the Subversion repository.

The release:perform stage just uploads the files on the EulerGUI Maven depot on sf.net .

One can do the unzipping through the shell on Sourceforge. But it is clean to do it with:

2009-08-24

Conzilla - A Concept Browser

Links: graph transforms, etc

Applying rules to form generation from OWL models

http://www.google.fr/search?hl=fr&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&hs=38t&q=grap%C4%A5+transform+logic+rules&btnG=Rechercher&meta=

Temporal Graph Queries to Support Software Evolution

Towards Model Transformation with TXL

From Graph to GUI: Displaying RDF Data from the Web with Arago

Towards Graph Transformation based Generation of Visual Editors using Eclipse

A Mapping Language from Models to DI Diagrams

Refactoring of UML models using AGG

Code Generation Network - Generators for MDA

Tiger Project

Conzilla - A Concept Browser | Overview / Conzilla - The Concept Browser browse

EulerGUI mailing list management at sf.net

N3 refactoring: extract part of antecedent

You want to reuse that rule with the rest of antecedent

Tn+1 ...
Tp

and the consequent unchanged, but with a different begining of antecedent.

You are tempted to copy and paste the original rule, and change the begining of antecedent.

In situations where the original rule OR would trigger, OR" then OR' will trigger.

The original begining of antecedent is in blue, and the rewriting with new predicates is in green.

Extraction of the rest of antecedent as a second intermediary rule (note that the consequent is copied verbatim ) :

Déductions generated applications : choice of a subclass

Use case : When the user clicks to choose among subclasses, the fields for chosen subclass are added at once, without fields for its superclass of course.

Problem with Drools recompilation

2009-08-21

Déductions generated applications : choice of a subclass

The text field and the focus lost listener both come from the same reason: a property value must be updated by the user.

2009-08-20

Rules about rules with Euler / Eye

This time it is an undoubtful error. I created a simple rule to infer RDFS properties from rules :

Déductions generated applications : choice of a subclass

2009-08-19

Déductions generated applications : choice of a subclass

DONE - For the RDFS food example, there is no more an empty GUI produced with editedClass == PreparedFood; this was because there is no properties at all. PreparedFood, however, has many subclasses. We give a choice to create and edit one of these subclasses. This reuses what is done for object properties.

Memory analyser

Maven eclipse integration

I can debug a single test run from Maven by creating in "Debug configurations" a new configuration under "Maven build"

N3 syntax for jEdit

2009-08-15

RDFS to OWL conversion - food example

TODO - I undersand why there is still an empty GUI produced with editedClass == PreparedFood; this is because there is no properties at all. However, it works fine with the CV RDFS model. PreparedFood, however, has many subclasses. We should give a choice to create and edit one of these subclasses. This is already what is done for object properties.

2009-08-14

Déductions generated applications

2009-08-13

2009-08-08

Article for OWLED 2009

FOAF: when owl:Thing is specified as a placeholder

The explanation is simply that "personal mailbox" is an RDF resource, and not a string, e.g. :

What is lacking in the FOAF specification is the information that this RDF resource is a mailto: URL (to be able to assist user input). So we should somehow enforce in the GUI the constraint that a the URL begins with mailto: .

The other interesting thing is that mbox is an Inverse Functional Property. The consequence for the GUI is that, given the Agent being the domain of the property, and in the absence of an existing mbox, the mbox is necessarily a new RDF resource.

2009-08-05

http://marbles.sourceforge.net/ Marbles is a server-side application that formats Semantic Web content for XHTML clients using Fresnel lenses and formats. Colored dots are used to correlate the origin of displayed data with a list of data sources, hence the name.

Eclipse [3] is an open platform for tool integration managed by an open community written in Java [19].

Déductions generated applications

2009-08-04

Déductions generated applications

2009-08-03

Déductions generated applications

2009-08-02

Generate readable diagrams

One must be able to generate readable diagrams, and that in itself is a big challenge. There is GraphViz occupying this territory; but what it does without specific instructions it is not beautiful.

GraphViz is already added to EulerGUI. But I do not give specific instructions, just the nodes and links

This is a nice use case for rule based processing; for a long time I have been running around this problem. My boss at MW was also interested in that. Now I am better prepared because I got like a training with the generation of forms.

Until today I had not thought of generating instructions for GraphViz. The least I can do is :

Also it is very important to connect specific diagram rules with pre-canned queries like:

Déductions generated applications

2009-07-31

Déductions generated applications

2009-07-30

Déductions generated applications

2009-07-29

2009-07-28

Conversation avec Alain Couder, PDG de Solid Information Technology - 14:56 - May 30, 2006

The classic Canadian documentary Manufacturing Consent based on the Noam Chomsky/Edward Herman book by the same name. Explores the the propaganda model of the media.

2009-07-27

Déductions generated applications

2009-07-26

intelligent modularity: example of graphic file formats; for e.g. OCR software one wants to seamlessly integrate any graphic file format ( which not the case of many graphic programs).

2009-07-24

2009-07-23

Misc. link on RDF, OWL

A Semantic Web Primer for Object-Oriented Software Developers : http://www.w3.org/TR/sw-oosd-primer/

XForms

The second part of the essence of XForms is that the form controls, rather than expressing how they should look (radio buttons, menu, etc), express their intent ("this control selects one value from a list").

Bind an XPath nodeset or ref to "facets" : type, constraint, required, relevant, calculate

2009-07-21

Déductions generated applications

Idea: JCHR

Instead of Drools, JCHR could be used for inferring forms in Déductions generated applications. But there no examples with Java user classes.

Update

Now the events are correctly passed to the rules storing it in the KB, at least for the first event (when no existing value for this property about this resource).

2009-07-20

Idea : intelligent window manager

Add Fuxi inference engine in EulerGUI

Trying CHR as a production rule system

2009-07-17

2009-07-16

TODO

Déduction generated application new architecture

2009-07-15

Hosted XML Database Solution ? We provide supported hosting of eXist XML Database!

OWLED 2009 - OWL: Experiences and Directions

2009-07-14

Links : B method, etc

2009-07-06

Generating XML from RDF

RETE engines in Phyton

Drools N3 engine: Duplicated triples

The "Person" test passes , but in some cases it might remove some correct results.

Anyway this a major improvement; it will make thingd easzier for Déductions new architecture.

2009-07-04

CALO, Cognitive Assistant that Learns and Organizes

Although neither application (under Windows only) nor source is available, some ideas could be extracted.

The tones project

The Ontology Modularization tool implemented an algorithm for extracting modules from OWL ontologies.

OWL 2

Each OWL feature is implemented in a separated file, generally with a single N3 implication. This allows to add to the rule base just the needed features. This makes for cleaner and more efficient rule bases.

2009-07-03

Drools N3 engine

What

To add a condition in the generated Drools code, I must have treated all the triples in the consequence, and is not convenient with my event based API.

How

2009-07-01

2009-06-29

The Sourcerer project

Inference engine with database access

- connect to Sesame database; this includes extracting the relevant triples from the database into the in-memory inference engines; both in EulerGUI and in the generated application.

But I realize that analysing a rule base to find a low volume extraction (from the most specific antecedents) from an RDF database is not an easy task.

A scenario both easy to implement, and that makes sense from a user point of view, is to add to the facts made available to the inference engine the result of a user SPARQL query from a SPARQL database. Here are the databases I selected:

2009-06-28

Explanations on N3 + Java in Déductions generated applications

Building Protégé from sources

That was not the right way; in fact there is no 64 bits version for GWT ! The explaination:

In fact, instead of compiling the C and C++ for 64 bits, I need to use a 32 bits JVM :( .

Proposals for the Protégé team

Use RDF/turtle format instead of XML or Java properties files

Leverage on semantics and rules for Java insfrastructure

In order to unify the GUI models for Protégé 4 Swing GUI and Protégé 3 Web, I should come with an architecture based on rules and a unified model for Swing and GWT (Google Web Toolkit), including the callbacks.

To ease the porting of Protégé 3 plugins to Protégé 4, maybe also an ontology based approach could be appropriate.

Compare Euler Eye engine and Drools N3

2009-06-27

TODO Déduction form generation new architecture

Trying Sindice

Sindice is a nice Semantic Web search engine. It is different from Swoogle in that it searches and shows fact triples, not ontology documents. There is a nice "advanced" query where one can enter e.g.:

Back from Protégé conference 2009 in Amsterdam

2009-06-21

UML Textual (UMLT)

2009-06-20

Tried opencalais

This web interface, from a NL sentence, marks it with RDF statements for the words it recognized.

Open Mind Common Sense

2009-06-18

Readings about logic

2009-06-17

Déductions app. generator: calling user Java code

2009-06-16

Ontology links

Recovered excellent presentation by Adam Pease "tutorial - Introduction to Ontology / SUMO"

2009-06-13

The OW2 Consortium is an open source community committed to making available to everyone the best and most reliable middleware technology.

A Programmer's Introduction to Pellet: How to Build Ontology-based Semantic Applications

2009-06-12

2009-06-10

Twitter message

Working on my next seminar on software engineering rules engines, and ontologies.

2009-06-08

Trying Protégé 4.0.111 with an N3 file

I didn't realize that Protégé is accepting N3 files ( I kept converting my files to RDF using CWM ).

Déductions: mixing languages in rules

This would be a good feature to reassure new users, although in general mixing languages is a bad thing.

e-commerce demo.

My first FOAF description

I rapidly created my first FOAF description jmv.rdf with foaf-a-matic by Leigh Dodds.

2009-06-04

Déductions: testing new generated GUI

<http://test#id1> <http://jmvanel.free.fr/ontology/examples/person#name> "Lennon" .

I don't have a handle, but I have an object that is equal() to the one I want to retract.

2009-06-02

Sound is silent on Ubuntu

2009-06-01

CouchDB

Idea: Using Déduction Forms for creating a Déduction application

Software categories

Is there any standard way (namespace, mapping of concepts, ...) of using the WordNet database in RDF statements?

2009-05-27

Testing APE (ACE, Attempto)

With the command line

With the Web client

Oslo

Conferences

Making invitations for his seminars in Paris

Changing status on LinkedIn

Jean-Marc is seeking what business can be done with his Semantic Web + rule based programming techniques

2009-05-24

Misc. links

BASE, BigTable, Dynamo(), Cassandra, Tokyo Cabinet, Redis, CacheDB, MongoDB(C++), MonetDB(bad), LucidDB(?), Vertica(C Store), FastBit, Bloom Filter,

Guice dependency injection

Updated EulerGUI summary and categories

Compare modelization OWL UML

2009-05-23

Tidying HTML

I used some HTML generator for images, and the HTML was not correct, so used the good old tidy :

2009-05-21

Uploading stuff to Maven central repository

2009-05-20

Setting time on Ubuntu wth ntpdate

Hyperlinks towards EulerGUI

Misc. links

http://www.dajobe.org/talks/200805-semtech/Yahoo%20Semantic%20Web%20in%20Production.pdf

2009-05-19

RDF databases and applications built on them

Groupware

Bug tracking systems

2009-05-14

2009-05-12

Input file for EulerGUI

It would not difficult to to write a rule set to populate the current class Project from a N3 document.: instantiate Project, instantiate the sources Map<String, N3Source>, add the URI to N3Source, etc.

A new file extension is needed, to avoid the confusion witth ordinary N3 sources. Or maybe just *.project.n3 .

Déductions: reflexions on multi-platform generation

This may add an exessive CPU time. There should be a way to "compile" the rule base given that we are not interested in intermediary results, but just by final Java objects.

Similar rules at instance level for properties of the translated objects; they are also simpler:

2009-05-08

Tasks

2009-05-05

Challenges

2009-05-04

Jean-Marc extends the Deductions frameworks for GUI generation to inheritance and adding pointers to a selection of objects.

2009-05-03

Deductions

Command-line User Interface

I think about doing a command-line User Interface (CL-UI) generator alongside the GUI generator. This will generalize the design, and make it easier to record user actions and explore intelligent behaviors.

Design GUI - KB architecture

After writing "Summary for AI researchers", and exchanging mails with David Holz, I'm ready to design the exchanges between GUI and KB . There are 2 issues:

I want the KB (Knowledge Base) to be dynamic and persistent, it's not only one inference pass and goodbye. It is necessary to infer a GUI, which is dynamic.

Proposed change: dynamic KB

Migration path

Tests

2009-05-02

2009-04-30

http://www.linkedin.com/groupAnswers?viewQuestionAndAnswers=&discussionID=3015442&gid=134407&trk=EML_anet_qa_ttle-cDhOon0JumNFomgJt7dBpSBA

http://www.lemonde.fr/technologies/article/2009/04/24/l-internet-des-objets-n-est-pas-celui-que-vous-croyez_1185201_651865.html#ens_id=1150782

2009-04-29

2009-04-28

Problems with Maven release:perform

I have the impression that, after downloading to target/checkout/eulergui, for some reason it doesn't execute goal 'deploy' here .

2009-04-27

Using Abbot on EulerGUI

2009-04-26

UMLT == UML Textual

EulerGUI demo. : Facebook rules

Next thing on my plate, integrate Sesame or OWLIM as a storage back-end, with management of the working memory extracted from the RDF store.

EulerGUI release

2009-04-25

Sesame first steps

N3 rule to change a literal type ?

I'm writing a rule base to convert eclipse eCore models to OWL; this works much like RDFS to OWL and UML to OWL . In the test input file, I have this:

aaai08 proceedings available !

Trying OpenCyc OWL Web interface

2009-04-09

Déductions

New use case

Beside generating from models, Déduction can bring rule-based programming to existing applications, by wrapping the Drools rule engine within Semantic Web languages. Many other use cases are envisioned, see below "vision".

Developers' training

2009-04-08

Trying Twinkle SPARQL GUI on DBPedia

Just download Twinkle , choose dbpedia.org in "Remote services" pulldown menu on the left, and paste the following SPARQL query :

morla rdf

Tried to install Morla RDF, but that doesn't compile on KUbuntu :( . Sent a mail to the list.

2009-04-07

Integrate ACE (Attempto) in EulerGUI

APE is available under the LGPL license, and it can also be used through a REST webservice. For example, an OWL ontology that corresponds to the ACE text

Slides presented at TAO Industrial Workshop

The slides presented at our last TAO Industrial Workshop that took place at ATOS, Paris La Défense on the 27th of January 2009.

All the video demos will be soon available on the Tao project website : http://www.tao-project.eu and you will also be able to watch the presentations on the videolectures.net website.

2009-04-04

Discussions on OO (Object Orientation) versus DL (Description Logics)

Added new mailing list for Déductions

2009-04-01

Connect on SourceForge with ssh

2009-03-29

Testing RDFS to OWL conversion

Having the OWL properties plus the RDFS properties in the same KB apparently is the reason for infinite loops.

It's too much trouble to filter out the unwanted stuff; it's really time for implementing the N3 processor pipes.

2009-03-27

Public repository for reusable models ?

By chance, is there a public repository for reusable Drools rulesets and models ?

I think one of the reasons for non-existence is lack of reusable exchange format.

UML is too complex, also unstable, and the strictly OO part (Class diagrams) is mixed with the rest of the huge 250 metaclasses metamodel.

eCore could be a good exchange format, but it's rather limited to the eclipse - Java world.

It seems that the best exchange format for business model is OWL - Web Ontology Language ( http://www.w3.org/2004/OWL/ ).

OWL has classes and properties . It has no methods and interfaces, but for that it is quite possible to reuse exactly the UML metamodel.

An advantage of OWL (and more generally Description Logics frameworks) is that a class defined by its properties (e.g. an UnemployedPerson is a Person that has no Job) provides a more natural way of modelizing a domain than classic OO, where you have to use delegation for the situation when a Person becomes UnemployedPerson.

And, contrary to classic OO models, there are sites for searching and hosting OWL and RDFS ontologies on the Web :

2009-03-26

Maven extra sources and eclipse

We generate the eclipse configuration with Maven, but I found no way to specify these extra Java sources so that they appear in the eclipse configuration.

2009-03-24

When searching for "RDF OWL Drools", I saw Mark Protor's blog on declarative-relational-programming , also speaking of RDF :

Searching for a generic XMI parser

If you saw my code, you would see that I'm traversing the UML model just like it were a plain eCore model (and by the way, I'm as much critic as you regarding UML, I heard you in Paris this winter).

So, what I'm asking is : is there a way to instantiate plain eCore objects from an XMI, disregarding the actual package ?

From what you answered I'm afraid that the answer is no, and that my use case was never considered.

Maybe it's enough to redefine this method in org.eclipse.emf.ecore.xmi.impl.XMLHandler :

Before being independent like now, I spent 2 years building an EMF-like framework with generator in Java, generating C++. And I did exactly that, an XMI parser that instantiates plain vanilla MOF classes. But I doubt if can use my work for an OSS project.

2009-03-23

http://en.wikipedia.org/wiki/List_of_UML_tools#Comparison_of_non-proprietary_UML_tools

Idea for rules applied to UML

Problems with Maven repo. for eclipse

http://www.nabble.com/working-around-eclipse-artifacts-in-repo-version-problems:-Couldn%27t-find-a-version-in--1.0.0-v20070606--to-match-range--1.0.0,2.0.0%29-td20054852.html

http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html

https://maven.atlassian.com/browse/org.eclipse.emf.ecore/xmi/2.3.0-v200706262000

XMI to N3 translator

This does not give a uml:Property object; there is only a uml:ownedAttribute predicate :

generic emf factory for any URI

UML input integration in EulerGUI

I'm working on UML integration in EulerGUI , by using the EMF eclipse infrastructure.

It works already quite well, with also rules to translate UML expressed in N3 into OWL (in eulergui/examples ) .

The only point is jar dependencies; I want to avoid adding UML2 jars, and also eclipse stuff on Maven central depot is bugged.

So I'm asking you if it bothers you if I temporarily add the modifications in EulerGUI SVN .

It will not disturb your runs, it's just that you will have one class not compiling.

To avoid having a large release, I'm thinking of using Java Web Start (JNLP) , plus the Maven JNLP plugin :

Hopefully this would allow to download the auxiliary jars on demand, when the user starts using the feature.

2009-03-21

UI designers : Untangle the knots, use EMF (live) models ! Olivier Moises - http://www.eclipsecon.org/2009/sessions?id=475

XMI to N3 Converter

ecore_raw.n3 is produced from a simple UML XMI file default.uml (produced by eclipse UML2 editor).

Then ecore_raw.n3 is simplifyed by the CWM utility to produce a more readable ecore.n3 .

EulerGUI Demo.

to add to Demo: show annotation of an existing resource that interacts existing rules: e.g. "good customer" => "send reward"

Reading and thinking on EMF

2009-03-20

2009-03-19

http://sourceforge.net/mailarchive/forum.php?forum_name=yap-users&max_rows=25&style=nested&viewmonth=200903

http://groups.google.com/group/fuxi-discussion/browse_thread/thread/d0db3434cc0c5e17

Déduction architecture

and plan.n3. The (specific) N3 rules to generate the HTML are here: plan-rules.n3 .

Design Java - N3 mixin

update RDF database after user change: keep same features as :update, but do it as much as possible with rule techniques.

Tried Netbeans profiler on Parser4J ( NetBeans eclipse comparison BTW)

Eclipse's TPTP didn't want to start (Unable to connect to the Agent Controller - Connexion refused). So I installed NetBeans 6.5.1 . Good news is that NB imports easily eclipse projects (using same directory possibly), inclluding run configurations. Bad news is that NB imposes more structure to a project that eclipse:

Because of point 1 , I couldn't compile the tests, so I removed them from the project. This dependency enforcing feature can be a very good thing, but it should be possible to disactivate it, especially when importing eclipse projects.

The profiler works well ; here is the result for Parser4J . It is really like stairs with even steps, so there is nothing that draws particular attention.

2009-03-18

Mixed paradigm architecture for Déductions updated

Building IsaViz from sources (RDF editor)

This means that typically, to checkout the foo/bar CVS Module, you would run the following commands:

2009-03-16

Installing eclipse 3.5 M6

2009-03-16

Development plan and status for Déductions updated

2009-03-15

Lazy load of jars - Maven ?

In addition to use Maven to manage dependencies and build my application (http://eulergui.sourceforge.net/), I would like to re-use the dependencies specified in the pom.xml to carry out a lazy loading of the jars in my application. That would make it possible to release my application with the most important dependencies in XXX-jar-with-dependencies.jar, and the application would download the dependences as needed, like Maven does.

I would like to avoid a big distribution with many jars, and a run.sh and a run.bat. Thus I would not like to develop myself this kind of infrastructure.

Thanks Michael; indeed it helps; I always thought that JNPL donwloaded a monolythic application, but that's not the case :

The JNLP file can list resources as "lazy", which informs the JNLPclient that the application does not need those resources to start,but can retrieve them later on when/if the application requests them.

2009-03-14

Déductions : debug inheritance in forms

Apart from my current problems ( the buttons for sub-classes don't show ) I should do this :

2009-03-13

2009-03-12

Ideas for Euler GUI

Misc. : pres. on rule-based programming, Science Commons, etc

jmv> Hi ! Is there a presentation showing the advantages of rule-based programming , preferably for developers ?

<mic_hat> jmv: yes there is a funny one involving pirates, will see if I can find it

Science Commons, a division of Creative Commons, is incubating the adoption of semantic scientific publishing through creation of a robust database of ontologies (http://neurocommons.org) and development of supporting technical standards and code.

what Xmarks is doing is very similar to what StumbleUpon's browser extension does too. : Evri, a semantic search engine

CAD/CAPP Integration using Feature Ontology ( CAPP = computer-aided process planning )

http://www.google.fr/search?q=drools+conseil+services+France&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a

2009-03-11

Tests with Drools engine + new N3 parser

2009-03-10

Déductions generated GUI

Integrating new N3 parser with Drools engine

Integrity Constraints

I come across a problem I avoided until now: integrity Constraints. This is what Euler EYE generates now:

It will be easy to report something also with Drools, although not as compact as EYE, which does elegant variable substitution.

2009-03-09

Discussion with Pieter Van Gorp - grabats 2009

later on, three case studies will be selected and all international groups working on model transformation are then invited to solve those case studies

so basically you could propose a case study for which you think you can very well illustrate the strengths of your approach

Pieter: obviously, you should propose a well-balanced case study for any approach but you can always have some extension points that allow you to show off particular features :)

if you do participate in this initiative, it becomes part of my daily work to review your work at some point

moi: I feel that the interesting thing in my approach is that I'm not restricted to a technological family , although It's currently based on Web Semantics and Java . I want to interoperate with UML, eclipse EMF etc, SQL , and others . Because, I'm not a standard researcher, I ' am practionner who wants to offer useful stuff for many projects, and who wants to develop a business .

2009-03-06

Ideas : JoaFip, EMF, JSR94

If JoaFip were compatible with EMF generated code (and it may well be), it could provide a non intrusive persistence solution for EMF .

I thought about using the JSR94 API for calling Drools, but in fact the N3 Drools engine itself could implement JSR94 .

LinkedIn current status

Jean-Marc writes a presentation on EulerGUI and Deductions frameworks for GUI generation. 3 days ago -

Jean-Marc extends the Deductions frameworks for GUI generation to inheritance and adding pointers to a selection of objects.

2009-03-05

Challenges of Deduction project: calling Java

The client visit yesterday was not successful, because I failed to show the benefits of rule - based programming , and the demo. was not convincing enough.

I want to propose them a partnership were I will develop features they need most, like outputting PHP, replacing XSLT producing SVG, or develop business rules ( they offer services for civil engineering management ). I think that the biggest challenge is to come with a way of augmenting the generated application with code in traditional languages so that :

The first idea that comes to mind is, following the "Intelligent Modularity" principle, is to modelize the (Java) interfaces for the "augmentation code". Thus point 1. is easily realized by eclipse, provided that java source for interfaces is generated. Point 2. can be realized by some checking coherence between the interfaces and the way they are called.

Who says interface also says "factory", so there can be a singleton factory à la Spring.

Another ingredient are extension points à la eclipse to add functionalities in well-defined places. For instance if one want to add action buttons in a context of a property, like the GRAPH and REMOVE buttons of EulerGUI, there will be an extension point for that . The generated interface for N3 property pref:MyProperty could be :

Each button will implement this interface. The design environment will be responsible for generating on the fly such interfaces, plus the Java methods to navigate the object graph , probably following a JavaBeans pattern (to detail <<< ).

The runtime environment will take care of actually calling the extension classes.

The runtime environment will also accordingly generate the glue code to navigate the object graph . There must be also a way to update (CRUD) the Knowledge base, like there is today the ITripleStore interface.

These ideas are quite straightfoward. Is this the way to achieve a seamless integration on the OO and ontology paradigms ?

EulerGUI : Drools engine - findall

or this with an empty result imposed (2) , here to check non-existence of subclasses :

A Drools query is not suitable, since the result of findall must be reused at rules firing time.

2009-03-04

2009-03-02

EulerGUI Generated application

COMMITED: able to make a round-trip with user N3 file and see the modifications in the GUI after re-opening the file.

Test VNC - serveur Linux, client Windows UltraVNC

... the amount spent watching TV worldwide equals to creating 10.000 wikipedia projects each year...

Cette fonctionnalité vous permet de rendre un de vos PC accessible depuis Internet (pour des applications de type serveur Web ou FTP) tout en continuant de protéger les autres ordinateurs de votre réseau.

2009-03-01

Thought about application in domotics of the "comprehension without protocol" principle

Deductions: to do before wednesday demo.

EulerGUI: Current data binding partipating classes

2009-02-27

Deduction runtime: storage

2009-02-26

EulerGUI 1.0 announce on LinkedIn and Facebook; developed my social network; invited for a presentation on March 12th.

2009-02-25

Experience with N3 rules - Tim Berners-Lee, Dan Connolly, Eric Prud'homeaux, Yosi Scharf

2009-02-24

Proposal for a vocabulary for RDF projects for development tools

Announce of EulerGUI 1.0

Announce on news:comp.lang.java.announce

I am pleased to announce EulerGUI 1.0 , a development environment for Semantic Web technologies , written in Java , under LGPL licence .

N3 is a language equivalent to Semantic Web's RDF language, but much easier to read and write by hand.

Among many other vocabularies searchable on Swoogle search site (http://swoogle.umbc.edu/), there is OWL from W3C, (OWL Web Ontology Language) that allows to express in RDF, and thus in N3, rich models with classes, properties, inheritance and constraints.

So in summary, N3 is a user-friendly language allowing to express facts, class and properties, and rules.

- export all that project as a set of Drools packages, plus the facts in XMLEncoder format

You can also use standard Semantic Web tools such as Protégé and Swoop to edit and visualize RDF and OWL files. EulerGUI is perfectly complementary to these tools:

- Protégé allows to classify and validate an OWL ontology via the classic algorithms of the Description Logic;

- EulerGUI makes it possible to develop a rule base in N3 logic, and to export it for a Drools runtime, or Prolog Yap

Misc.

Ideas

2009-02-23

Problems with Maven release:prepare solved

<jmv> I don't if I should give the developerConnection and tagBase as local file: path or repository https: path .

<jmv> I understood the SVN message : "Local operations do not take a log message"

<jmv> So I tried specifying <developerConnection> and <tagBase> as https: URL's , not file: .

<jmv> NOTE: several pages give bad examples of specifying <developerConnection> and <tagBase> as file: URL : e.g. http://maven.apache.org/guides/mini/guide-releasing.html

<jmv> Also the tags directory on the repository is not created automatically by mvn relase:prepare , and nothing in the doc. says it should be done manually before .

<jmv> I am probably one of the rare people trying to do it the Maven way before having done a single release tagging before .

<jmv> I wonder if it's good practice to use http://MY_PROJECT.sourceforge.net/ as Maven depot ...

<jmv> I suppose that if I want a plain jar in addition to the jar-with-dependencies I have now, I just have to add :

<jmv> Well, it seems that I already have a plain jar here, without asking for it : http://eulergui.sourceforge.net/maven2/eulergui/eulergui/1.0-SNAPSHOT/

<jmv> However, it didn't upload the new release as eulergui-1.0 ; I guess I have to review my deploy specifcation ...

<jmv> I have another complaint : after regenerating my pom.xml upon release:prepare , it left it like this, and I'm afraid it will trigger the same problem next time :

<jmv> <developerConnection>scm:svn:file://PCIVENTIVE/home/jmv/src/eulergui_test/eulergui/trunk</developerConnection>

<handyande> jmv: prepare updates the pom, commits, branches and then updates the pom again

<jmv> handyande, That's fine , it's just that I had problems with initial pom.xml having <developerConnection>scm:svn:file: , and it worked with initial pom.xml having <developerConnection>scm:svn:https:

2009-02-22

Problems with maven-assembly-plugin + jar-with-dependencies

The reason was simply that a copy of this jar was, for an unknown reason, at the project root (not in target).

2009-02-19

2009-02-18

Drools bug ?

As I mailed to Edson Tirelli, I'd like to know whether loading rules and facts in any order is allowed (before firing rules) . Of course I'm about to change the way I call Drools, but before that I'd like to know the status of this feature, is it a bug or is it forbiden.

2009-02-17

On the path to a Drools bug

I extracted from a larger model 2 rules. The problem is that the results from rule "0" should include all the results from rule "1" .

Now, the issue is that in some test cases, large enough, there are results from rule "1" that are NOT included in the results from rule "0" .

I feel that the problem is related to the fact that criteria C1 and C2 are , in the particuliar case of the offending Triple item, the same object .

2009-02-12

I tried to make these stupid NTP infrastructure work, for keeping the machine up-to-date, on Ubuntu. No way ! Then I simply did this, looking from my watch:

Problem solved with Drools

I had to comment out the usage of findall, as it is not yet implemented in Drools translation ! And that was the main problem!

Now there are many such messages on NRI 2 (NRI with individuals removed), but the GUI looks good anyway :

2009-02-11

Testing APE

Debug with Drools

<jmv_> Hi all ; I sent a mail on the user list "debugging generated Drools source code", if you want to chat here, I'm available ...

<jmv_> To better read the logs , I should better understand what is a tuple and the fact inside ...

<etirelli> jvm_: for your case, audit log is better... I don't think that looking at the generated code will have any value for you...

<jmv_> audit is something different from what I have with DefaultAgendaEventListener . afterActivationFired() ?

<krisv> jmv_: the audit viewer extracts the most useful information for you and shows it in a tree view

<jmv_> My rule run lives inside an application in a non-Drool project . Is that preventing me from using the audit view ?

<krisv> no, the log is an xml file, you just need to open that with the audit view

Drools engine for N3 logic: Drools interrogations

After reordering results from Drools according to recency (probably useless), and correcting the Euler parser output (was introducing extra rules), I came to a more fundamental issue. To see that, I logged Drools events "After Activation Fired", and named rule by file name and number, e.g. rule=query.n3 0 .

it should unify with both #1 and #3 . This seems to be a problem with Drools, I need to clarify. As a workaround, I added this rule. It is clearly redundant with a logical engine, but with Drools I'm not sure. But in one case, it made things better (NRI reduced).

Now when I test a small case in isolation, it works, which means that Drools can accept a single instance.

So I'm puzzled, Drools seems OK, but my translation doesn't work on large cases.

At least my rules are not suspect! Because they work, although slowly, with the Euler Prolog engine. But the translation is suspect, and so is Drools ...

What is a systematic of debugging a rule base that doesn't trigger when expected ?

2009-02-08

Drools engine for N3 logic

I thought I needed to reorder results from Drools according to rule order and fact order, but there were other problems.

2009-02-07

Workshop on Controlled Natural Language (CNL) 2009

I would like to attend that. It's free, but would I be admitted? Mmy stuff is quite remote from that now.

EulerGUI project

Drools engine for N3 logic

Drools N3 engine: implemented fn:substring-after, at least so that case "order request" works almost as with Euler engine.

For the big NRI test case now we have a few seconds before getting the GUI, instead of hundreds of seconds with Euler-EYE !

The remaining root problem for the big NRI test case now is still to respect original order of properties in OWL:

Use N3 format ( TODO )

Using Instanciator class, it would not be difficult to read an N3 project file and translate that into n3_project.Project file. But the reverse operation is not so obvious, when one wants to avoid ad-hoc code. Some way of converting a JavaBean into RDF is needed ...

2009-02-06

SQL humour

Drools engine for N3 logic applied in "Déductions" GUI generation

Current situation

After some laborious work, I get something working with the Drools engine almost like the Euler-EYE engine. The Actions for text fields and buttons work, because the RDF list in Consequent is correctly implemented. The log:uri built-in is implemented, and the numeric comparison, at least in some call modes.

There is a Drools Working Memory that stores the end-user data, but it's not visible since there is no load / save in the end-user (generated) application.

Missing features and plan

There is a common theme in the 2 items above; instead of hand-made procedural code, we should use a declarative rule-based engine, just like in "Déductions" for GUI generation. Point 1 is important because this is the embryo of a multi-language, multi-platform tool. Point 2 is more complex because this is a bootstrap issue: a rule engine is needed to build a rule engine.

With the current Drools-N3 engine, there cannot be more than one value for a subject and a predicate. I don't think this is a severe limitation.

Redesign the JavaScript+Java instanciation code

The current class is Instanciator. When processing RDF triples, we want basically to say:

This looks pretty simple! But it takes 646 low-level Java lines, specialized for a specific naming convention in input, and JavaScript+Java object on output. We must be much more flexible in input and output.

What is an instantiation in most languages ? Well there 2 cases: typed languages and weakly- or non-typed. The ingredients are:

I think of using a RETE engine, but I don't like Drools' syntax. Think of the number of pages it takes to describes the stuff ! They are stuck in an old variable instantiation-variable use scheme. Compare that to the flexibility of Prolog! So I'd like to try another engine, and/or wrap the OO RETE features into an N3 or Prolog-like vocabulary or language.

The way Hammurapi expresses rules is interesting: in ordinary in Java language inside Rule.insert() methods. But it's not what I'm lloking for. Reading Hammurapi made me thing of using JSR94 API instead of plain Drools API.

2009-02-05

Drools

drools-docs/drools-docs-expert/src/main/docbook/en-US/Chapter-Rule_Language/Section-Rule.xml

2009-02-04

Prolog SWI - Yap compatibility

First I replaced numbervars with number_vars everywhere, because this is a predefined predicate on SWI but not on Yap . Then these two conditional blocs do the trick:

Now I have something that compiles on both SWI Prolog and Yap, and I can use the superb SWI Prolog IDE. For more complex problems, I advise to read this paragraph in SWI manual;

2009-02-03

Debugging my rules for GUI construction

And the result on NRI, with expandForms disactivated, and the comprehension limit raised to 600, is :

With expand sub Forms activated , there was no problems with cycles in the class-property graph, but the CPU time raises now to 1799 seconds!

Note that the form with expand sub Forms is just for demonstration purposes; it has no scroll bars and is larger than the screen.

I tried yesterday's point 4, by adding this global settings that is false by default:

I thought that the app:isReachableFrom relation is expensive to compute. Applying that "enhancement", we go from 7s to 116s on the "order request" case. The reason is probably that we can't eliminate the support classes when doing this:

2009-02-02

Debugging my rules for GUI construction

Tests

When watching the run with --debug, I saw that much time is spent on navigating the relation between class and property and the class hierarchy .

Computing the transitive javapr:contains relation , transitive closure of javam:add is useless , as this was added for testing purposes. It's the same for gui:contains . I don't don't why they get computed at all; I thought that the Euler-SEM algorithm was goal directed. Anyway this is probably not the root cause of the infinite loop, but removing the unused stuff may help to find the problem. In fact the CPU on test_person went from 20s to 15s by removing these 2 relations.

At one time I thought that a cyclic relation between class and property was the problem, like this. But this runs fine.

Ideas and remedies

To implement point 3 (split the inference into steps), I need a way to query the intermediary results. The query used in current Déductions Project is expensive, as it uses inheritance on the set of Java classes used:

for a single ?OBJECT, so we need to filter out the abstract classes manually in the N3 to Java translation, and also rely on these triples to come out of Euler in the right order.

There is a simpler way to query, for the intermediary or final results. Simply add a marker property meaning : "this object belongs to the results" . So the query would be :

I can also make use of scoped computation of most Derived Classes for a given instance.

2009-02-01

Test OWL Inconsistent ontology

I have problem with the NRI SQL model that passed though RDBToOnto converter . The tool chain fails:

SQL ==> N3 ==> my rules for GUI construction with Euler ==> error(comprehension limit exceeded) ! :(

Indeed the original SQL has INSERT INTO declarations, that translate in RDF without datatype indication, e.g. :

Then I run again RDBToOnto converter, this time specifying no instance , and then Pellet says :

So I go back to suspect my rules, or rather a combination of my rules plus this 16 classes, 32 objects and 59 data properties OWL model.

Debugging my rules for GUI construction

It must be some longer cyclic implications involving my rules, plus property domains and ranges, as there is no sub-classes at all.

I began to play with queries on the rules themselves, e.g. this rule detects the rules having a triple with property javapr:projection in the consequent :

I could also by dichotomy remove rules until error dispappears. My best hope is to trace reasonning steps.

2009-01-31

Tidy EulerGUI

2009-01-29

Déductions - GUI generation : popup frame for object properties

The trick is to use the forms computed at design time by Euler as prototypes that are cloned at run-time. The ObjectBehavior class maintains a correspondence Map between GUI Container and RDF Object Id , which allows to keep track of the instance edited by the GUI component, without modifying the latter.

This is quite fragile and roughly acceptable for demo. purposes; the right implementation will be to use Drools engine at run-time.

Trying a large OWL with Déductions GUI generating rules

I took my former colleague Thomas Cenni's NRI-DB medical database . I passed it though RDBToOnto converter (a very nice tool BTW) , got a 2600 lines N3 - OWL file. Yap Prolog still runs now, after 360 minutes !

Converting my local N3 project (EulerGUI project) into a equivalent project referencing the SF.NET Subversion URL's :

2009-01-28

Déductions - GUI generation : data binding

The Java components I create currently for forms are in fact prototype objects. A single form can be instanciated several times simultaneously, for multi-valued fields. They should not be shown inconditionally at startup time. Instead, when user clicks on a button for an Object Property, the forms are instanciated and embedded in current frame or in a popup window. In java Swing, a Component cannot be embeded in two different containers.

At runtime, a form or an input field instance is associated to an edited instance, which corresponds to a triple in RDF. But we are editing a specific instance of triple object in the triple store, and we are modifying it in place.

The simplest would be to have an inference engine at runtime. But we don't have one right now (and also infering a form every time a user clicks a "create" button might be too slow). So a flexible design is needed here.

Euler GUI

Euler GUI: prepare for next release: avoid installing Eluer: fallback: copy euler.yap and codd.properties from classpath to temporary files

Links: AI

2009-01-25

N3 block comment

2009-01-23

Déductions - development plan

Experimenting with e:findall in Euler

I realize that e:findall (which is a wrapper around the ISO Prolog predicate findall), doesn't work like this:

The variable ?TYPE must be bound somehow for this to work (bug or feature?). For instance with gui_generic.n3 from the Déductions ontology I can add this in the antecedent :

I am surprised by the empty {}, but that is not a problem, I can replace true in the consequent with 1 .

Note that this is just an example. In the actual rules of the Déductions ontology I need to find the most derived classes of business (user) classes like purchaseorder.owl; the aim is to make GUI buttons to create instances of the most Derived Classes.

Using (?SUBTYPES) for the resultset was enforcing that there should only be one result item.

Open an N3 ontology in Protégé 4

I find Protégé convenient for exploring and visualizing an ontology (but not for writing one, were I prefer to write N3 source). First, convert the N3 into RDF:

Note that this would be a convenient feature in EulerGUI: adding a button to export an N3 file as RDF. Then the resulting RDF file can be opened in Protégé:

2009-01-22

Euler suggestions

Wouldn't it be possible to create a special N3 predicate for debug trace printing that could be added in antecedent or consequent ? Something like :

where options could be empty or "print_all_variables", "print(?MY_VARIABLE)", etc . When empty, it could print a count of times called, and maybe some indication of the current state, like current CPU time, number of predicate called, etc.

In this line of thought , I'd like to have a note on how to debug N3 rule bases; I could write a first draft, and you could add stuff from your vast experience.

2009-01-21

Customer with existing Access application

I'll visit a customer friday, to enhance a knowledge keeping application based on Access. While thinking on ways to reuse existing VB code, I thought about the Mono project , that could run VB code on Open Software platforms. BTW Mono could be one of the back-end platforms for Déductions. I don't know enough of the Access VB dialect. Does it use standard portable API like ADO.NET ? Well, the answer is that it can use one or more of several API's that Microsoft did over the years : Microsoft Access Object Library, several flavors of DAO (Data Access Object), ADO.NET . See MS database API's, http://en.wikipedia.org/wiki/Visual_Basic_for_Applications

2009-01-20

Déductions - GUI generation : work on subclasses

If the Property type has any subclasses, only the most derived class(es) must be exposed to the user as explained in paragraph "Create or link". If not, the Property type itself is exposed to the user.

EulerGUI - documentation

I updated the documentation with install and usage of latest Euler EYE processor (with 100% Prolog engine replacing old Java N3 to Prolog translator),

Errors in Euler / EYE (new stand-alone Yap engine)

Jos corrected the first error, not the second one. He also corrected the missing object in triple error.

I implemented the strings delimited by """ for Java code generation. So now I can really work with the new EYE engine.

2009-01-19

EulerGUI - integrated new Prolog self sufficient engine

- Euler installation directory is searched in expected places (EULER_INSTALLDIR)

And on deductions/n3/euler_project.xml.n3p , there is a different error : an object part is missing in an triple at third line :

Déductions - GUI generation : work on subclasses

Classes having subclasses used to work (in example deductions/n3/euler_project.xml.n3p ), but I broke that when working on purchase order. Now I need a way to say : "when a Class has subclasses , do this with the most derived classes, and otherwise do that with the Class itself ".

For this I'm trying to make use of "Scoped negation" in N3, namely log:notIncludes and e:findall with an empty result set.

2009-01-18

Déductions - GUI generation : implement Property details

Everything is expanded, for instance the "hasCustomer" button is followed by a corresponding "Customer" titled border (sub-form), etc . Soon we will connect the sub-forms to properties instances, and offer a choice of expanding sub-forms or not. And of course the "hasCustomer" button will create a new customer. However in this case "hasCustomer" is a one-valued mandatory property so behavior can be simplified, by expanding the sub-form unconditionally.

2009-01-16

Déductions - GUI generation : purchaseorder.owl

JDK7, FAM

Misc. confguration: NTP, SMTP

The Property - Class graph

2009-01-15

Déductions - GUI generation : splitting rules

What I did when splitting rules because of non mandatory rdf:label resulted in cleaner design.

Now I must do the same refactoring for Buttons for Object Properties, whose "text" property comes also from rdf:label.

2009-01-14

Déductions - GUI generation driven by logic rules

I had a look back at my Generic JavaBean editor included in ecolosim project . It has goals similar to GUI generation in Déductions project, except that in ecolosim the starting point is a set of Java classes respecting the JavaBean naming conventions, whereas in Déductions it starts from an OWL ontology. An important limitation of GUI generated in ecolosim is that properties of non-primitive types are always considered embedded (there is no way to create a linked data structure). In Déductions accurate GUI's will be inferred from the input ontology, taking in account all constraints (cardinality, linked or embedded, ordered or not, set or not, etc).

2009-01-13

Déductions - graph transform driven by logic rules

I think that what I have been doing for GUI generation is just a particular case of object graph transform. The limitation compared to XSLT or forward chaining rule engines (Drools) is that there is no rule priority, or default reasoning. This is a limitation, but also a cleaner way to do, since rules are both constructive and logical implications.

Using RDF ID when no rdf:label - Generated GUI - déductions

I solved the yesterday's problem , by using the RDF ID when there is no rdf:label (no other way).

Precisely, both elements in red are removed, and this triple is added in the consequent to keep track of the concrete newly created Label widget :

I found a good enough (for now) way to solve the default values issue. In app_gui-rules.n3 , putting the default rule (use RDF ID) before the specific rule (use rdf:label) , the JavaScript assignments are in the right order, override one another, and the resulting GUI is OK :

2009-01-11

Generated GUI - déductions

I realise that the Purchase Order exemple doesn't output any form field, just because the rules (app_gui-rules.n3) expect an rdf:label on each property. I should decouple the rules concerning label from the ones concerning input fields, or even better make a rule to enforce that undocumented widgets such as input widgets should be accompanied by a GUI label.

But the most important thing is that somehow I create a label out of the property Id , e.g. "has Customer" from hasCustomer here:

First this is not very clean to use an ID, and second we should avoid default reasonning, which is non monotonic logic. Anyway as far as I know, there is no way to do non monotonic reasoning with N3 logic.

EulerGUI - TODO - new window

A new project window must have an initial directory for file chooser taken from the current window.

2009-01-08

Deductions architecture

I'm thinking of using Drools engine for non-monotonic stuff, especially generating JavaScript code (see 2008-12-24).

Also to generate XHTML I would need some framework allowing to go from RDF repesentation to XML.

I'm thinking of using JRDF instead of my own simple Triple class . It would be good either to wrap a JRDF implementation so that it also stores on a Drools WorkingMemory , or completely rewrite an implementation of a JRDF interface in terms of a Drools WorkingMemory .

Or I can simply take inspiration on JRDF API, and basically keep my stuff for now. To implemeted my generated GUI, this thread is relevant :

2009-01-05

Looking at RDF Java API : jrdf

2008-12-24

Deductions architecture : generate other GUI's than Swing

We need to refactor java_projection-rules.n3 to create an additional layer with abstract widgets generalizing Swing and HTML.

This layer will be useful also for generating SWT with Wazaabi2 by Olivier Moïses, or generation of GWT, etc.

The straightforward solution in the line of what was done before is to replace strings like this in java_projection-rules.n3 :

And then rules must be added to translate widget:XXX into the kind of widgets for the plateform considered.

In the Java Swing case, this is particularly simple, since the name of unified widgets are just the Java Swing names with the initial J removed.

and besides it is not accurate, because for several widget:Button we will get a single java:JButton, unless we add this in the LHS :

with cardinality restriction 1 on javapr:projection, or describe it as an functional property .

2008-11-11

Translating N3 into Drools language

Here is a test I did with Drools before writing an automated transform for EulerGUI. One can see that N3 is more compact than Drools . An annoying thing is that implicit variable binding is not allowed, which make Drools less declarative than N3 and Prolog, and complicates the translation.

2008-11-06

Dialog on the Drools IRC

* |conan| (~conan@87.113.106.171.plusnet.pte-ag2.dyn.plus.net) a rejoint #drools

<jmv> instead of parsing a Drools source, I want to call methods to create rules, much like DOM in XML

<etirelli> jmv: you can use the classes we call Descriptor Model... you just need to be aware that although their are mature and rarely change from version to version, they are not considered public APIs and may have minor changes over time

<etirelli> jmv: http://anonsvn.jboss.org/repos/labs/labs/jbossrules/branches/4.0.x/drools-compiler/src/main/java/org/drools/lang/descr/

<etirelli> and examples are only the integration tests.. since it is not really a public API

<etirelli> to learn how it works, right sample rules in DRL, and use the DrlParser to parse them into descriptors... inspect the descriptors to see the generated result

<jmv> time permitting, there might be a Drools engine here : http://sourceforge.net/projects/eulergui/#

2008-11-02

Deductions project

Advanced JavaScript within JDK

In fact new Object() is not necessary; it is a plain JavaScript object, i.e. an associative array.

Note that in json2.js the JavaScript incantation to introspect the JavaScript object is :

2008-10-19

Deductions project

At last the Deductions project is on Sourceforge ! The Subversion repository is populated with N3 sources and the HTML that are on jmvanel.free.fr . There is a meeting on november 15th, and possible work with Olivier M. on an SWT backend.

The Deductions rule base can generate a crude window; but the multiplicity is not taken in account. And for now there is nothing for runtime behavior and storage in end user application. But it is declarative all the way :-) .

One of the goals is to remake the Euler GUI application with the Deduction technology.

2008-10-07

Deductions project

Reading N3 logics specification

Write non-regression tests

Componentization of software file search feature

Google Chrome

2008-08-24

Trying CWM proof checker

2008-07-07

Euler N3 reasoner

I developed a little Java class, Project, to shortcut the intermediary Web server and reuse the same rules and facts compiled in Prolog for several queries. It's a minimal file based wrapper for Euler SEM reasoning services; a Project owns a Set of N3 sources, N3 queries, and possibly Prolog sources.

On top of Project, I did a small Swing GUI. It's still barely usable, but it already can load several N3 sources and queries and lunch the inference process. I plan to add an RDF reader through CWM and Jython.

SWI Prolog

2008-05-19

moi: garder de l'information fermée sans sans que les utilisateurs puissent y accéder de manière neutre devrait être interdit ....

2008-05-08

2008-05-04

Building Protégé 4 from SVN

2008-05-01

AI links

Transformation d'ontologies basées sur la logique de description : Application dans le commerce électronique

http://tel.archives-ouvertes.fr/action/open_file.php?url=http://tel.archives-ouvertes.fr/docs/00/21/41/45/PDF/These_LeDuc.pdf&docid=214145

Description Logics for Conceptual Design, Information Access, and Ontology Integration: Research Trends (2003) - Enrico Franconi

Ian Horrocksa, E-mail The Corresponding Author, Birte Glimma, E-mail The Corresponding Author and Ulrike Sattler

http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B75H1-4NTXG8S-2&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=797599863fa17525cb0ea07f88e6b27f

Installing Coq on Fedora

Workflow for a DL + Prolog framework: trial of Euler

The Prolog rules are expressed with the same predicates and reified concepts as those translated from the initial DL (Description Logic) Ontology, and are the part of the knowledge that is too complex to be part of the DL Ontology. An example of a non-DL rule is here:

Questions on Euler

Notes

Questions

Q: The output of Euler is a proof in N3. Is there a way to shortcut that and produce a Prolog output ?

Q: Note that with the latest version of Euler and cwm this gives a parser error in cwm :

Q: How would you do to integrate Euler into a Java GUI such as SWOOP or Protégé ?

A: it depends on what one want to achieve: one feature could be to write an N3 rule and display the proof

Q: What is the difference between Euler and classical DL reasonners such as Pellet and FaCT++ ?

A: Euler is an all-purpose logical query engine. Reasonners process the ontology globally, thus achieving much better performances than processing the classes one by one.

Q: How can one use Euler to implement the services of a DL reasonner, i.e. classification and detection of unsatifiable classes?

Q: Before reading more about the Euler path algorithm, I wonder if using a tabling Prolog such as XSB would not achieve a similar effect, that is to avoid infinite loops and memoize the calls already encontered.

A: A We took the option of a reasoner written in Prolog instead of taking Prolog as a reasoner and there are many reasons for that: no need to declare which predicates should be tabled, having a proof output (in N3), supporting Coherent Logic, etc..

Q: Which crucial feature made you switch from Yap and tuProlog to SWI ? The semweb libraries ?

A: not yet; I would also consider to use jTrolog (which has better performance) https://jtrolog.dev.java.net/

Q: What are the advantages of the current SEM (Skolem Euler Machine) ? More expressive power or performances ?

A: It has more expressive power in that one can have existentials and disjunctions in the conclusion. This allows to find e:falseModel and e:counterModel but I still have to investigate the proof expression of those. For Horn Logic the engine speed is not degraded which was quite a challenging task :-)

Q: I tried to read the thesis mentioned on the site; it's not easy; also there is a different language ML or haskell .

Q: What is the difference with the N3 parser in CWM, apart apart from Euler being in Java ?

A: The Jena N3 parser can't cope with {} formulae which is crucial for N3 Full support.

The good servant: example

Example of stupid behavior in a browser (Firefox 3). I saved all tabs in a single bookmark, then I wanted to close the window. The browser, however, asked me to confirm that I really wanted to close the window.

2008-04-20

Trying tabled Prolog: XSB

2008-03-29

Trying Sigma Kif Web application

Programming advice: let the semantic be explicit

Representing weight in FOL

While I'm practicing on this exercice in AIMA I try to express in first order logic in an elegant way that :

∀s ∈S,

weight (s) \leq C \Rightarrow weight (bunchOf (S) \leq C . card (S)

Upper ontologies

From the Sumo site one can enter an English word and the corresponding formal term in SUMO for each WordNet sense (synset) will be displayed.

The open source Texai project has released the RDF-compatible content extracted from OpenCyc.

A bad sign: the research.cyc site is down this saturday. It's too bad, because on top of what Opencyc has, it includes, according to Wikipedia, a large lexicon, English parsing and generation tools, and Java based interfaces for knowledge editing and querying.

There is at least one exchange format for FOL (first order logic), the one used by the TPTP (Thousands of Problems for Theorem Provers) Problem Library, which is a library of test problems for Automated Theorem Proving (ATP) systems.

The most used format for ontologies including logic rules is KIF; generally an OWL extraction is provided; but OWL doesn't provide rules.

SUMO

2008-03-22

I'm practicing on this interesting exercice in AIMA, using Prover9. I won't put my solution here, thinking of teachers who are using this for home work, although I think that having a corpus of solutions is a good way to learn.

% 4. If Mary was buying tomatoes at the same time as John, did he see her? [Yes]

2008-03-19

OCL and First Order Logic

While searching for OCL + FOL, I found Isabelle, which is a kind of generic logic tool (not only FOL, but Higher Order Logic, etc).

Prover9 : how to recover the answers

My (JMV) comment: you can't use the answer "attribute" in unrestricted formulas, you have to convert then by hand to clausal form first, otherxwise you get this message:

Fatal error: Answer attributes on non-clausal formulas cannot contain
variables

2008-03-09

CHR

I subscribed to the CHR mailing list; there are quick responses to my newbie questions by the bosses themselves, Tom and Thom.

Blog versus HTML site

I updated to Amaya 10.1 on Fedora 7. It's quite convenient to write my Blog this way. Of course there is the hassle of uploading. But on Google JavaScript editor (mail, docs, or Blog) it tends to be quite slow, because of general slowness and memory consumption of Firefox.

2008-03-02

CHR, SQL+ XSB

Among other news, he started a type checker library in (SWI-)Prolog: type_check.pl. Using this library, you can use Mercury-style type declarations and predicate signatures. It's a migration path to Mercury, and also it's more flexible than Mercury, because typed Prolog code can call and be called by ordinary untyped Prolog code.

2008-02-17

CHR

2008-02-03

2008-01-28

Intentional programming; describing libraries formally in term of functionalities

I also read the Wikipedia page on Intentional Programming . It's clear that Intentional programming has part of the thruth. But also the product that personifies the concept is probably a big fat GUI with a proprietary file format, the triumph of proprietary software. Not what people want in 2007. They want their design efforts to be saved in an open and durable format. Guess what it is, dear reader ? OWL. There is no other. And people in 2007 want flexibility. They want to use today's tools and source code with emerging paradigms.

With a suitable inference engine and rules, plus a semantic search engine, plus a suitable GUI, this can enable to plug any annotated module to an exiting application. Remember that, like in eclipse, plugins are a composite structure, i.e. a module can have any granularity. So in theory this can solve the application building problem, in a top-down way. Of course, you'll still need develop features that exist nowhere. Also you'll still face the Babel effect on programming languages. But connecting different languages is "just" a matter of correspondance rules and, again, protocol state machines. So connecting different languages is automatizable; it's mainly an issue in performance and global architecture.

2008-01-27

My gnuplot tricks

2010-02-07
- List of Projects That Use Protege
- Gloze problems
  - The gloze.base parameter
  - Gloze within EulerGUI
    - The XML rood node is an RDF blank node
2010-02-05
- RDFa
2010-02-03
- To read
2010-02-02
- Haskell
- Recent EulerGUI commits
2010-02-01
- Recent EulerGUI commits
- A multi-valued property that is a LinkedHashMap: question on ##swing
2010-01-30
- Specification of EulerGUI Server : parametrized queries
2010-01-27
- Installed Thea
- Running FuXi from repository: trouble
- Links
2010-01-25
- Readings: ACE View, Jena 2 Inference support
2010-01-22
- Misc. links on OWL reasoners
- (Re)Tried Open Mind / Common sense
2010-01-21
- Recent EulerGUI commits
2010-01-20
- Announces EulerGUI 1.4.1
- EulerGUI bug: creating a new SPARQL query
2010-01-19
- (re)trying Euler server
2010-01-18
2010-01-13
- Trying Joseki from sources
- A server that wraps EulerGUI
- Updated Fuxi
- Subtleties of regular expressions
- Drools N3 engine: factorial test
2010-01-11
- RDF servers
2010-01-10
2010-01-08
- EulerGUI 1.4 Announce for English language followers
- EulerGUI 1.4 Announce on Protégé list
- Drools N3 engine: built-ins
2010-01-06
- Trying TDB Jena
- Intelligent test generation
- ordered lists in OWL
- Links: Non-monotonic_logic, Closed_world_assumption, etc
2010-01-05
- Attempto (ACE)
2010-01-04
- Releasing EulerGUI 1.4
- MVEL
- RIF (W3C Rule Interchange Format)
- Attempto (ACE): tried some examples
- Tests with empathy and Google Talk
- introductory material on the semantic web
- neo4j
2010-01-03
- Checking EulerGUI dependencies
- RuleML
  - RuleML and ACE
- Attempto and N3
2009-12-24
- Consolidating EulerGUI
2009-12-23
- Gloze
2009-12-22
- Gloze use cases
2009-12-21
- Gloze
2009-12-20
- Gloze
2009-12-19
- Gloze
2009-12-18
- Recompiling Gloze 1.01-beta2
- Debugging Maven test cases with eclipse
2009-12-17
- EulerGUI
2009-12-16
- EulerGUI
  - XML export with Gloze
2009-12-15
- Fetching Jena sources from depots
- EulerGUI
  - XML export with Gloze
2009-12-14
- EulerGUI Drools N3 engine
  - Cardinality
2009-12-13
- EulerGUI Drools N3 engine
  - Cardinality
2009-12-12
- Installed Drools eclipse plugin
- EulerGUI Drools N3 engine: cardinality
2009-12-10
- EulerGUI Drools N3 engine: cardinality
2009-12-09
- EulerGUI Drools N3 engine: cardinality
2009-12-08
2009-12-04
2009-12-03
- EulerGUI: SPARQL queries
  - Enhancement in opening buggy projects
2009-12-02
- EulerGUI : enhancements in tests
  - Bypass the Parser4J tests
- Disactivate the windows in Jemmy tests
- Testing my modifications in Gloze
2009-12-01
- Testing Gloze in EulerGUI
2009-11-30
2009-11-29
- EulerGUI + Déductions : update Jemmy 2 test
- Trying Jester, a test tester
- Joafip + Drools
2009-11-28
- EulerGUI + Déductions : update Jemmy 2 test
2009-11-27
- EulerGUI + Déductions : update Jemmy 2 test
2009-11-26
- EulerGUI + Déductions : update Jemmy 2 test
- On phone with Olivier Moïse
2009-11-25
- EulerGUI + Déductions : added a Jemmy 2 test
2009-11-24
- Looking for java gui testing frameworks
- JToolbar problem with uispec4j
2009-11-23
- My PS1 for zsh
2009-11-22
- IDE's and tools
- Questions about Abbot
- Working on Joafip + Drools
2009-11-21
- Working on Joafip
- Putting Abbot on Maven
2009-11-20
- EulerGUI TODO
- Rules for form generation
2009-11-18
- Rules for form generation
2009-11-17
- My Philips webcam works with cheese!
- Rules for form generation
2009-11-16
- Rules for form generation
2009-11-14
- Rules for form generation
2009-11-13
- Webcam on Packard Bell
- Rules for form generation
2009-11-12
- Aleph & Inductive Logic Programming
2009-11-05
- Explanations about rules for form generation
2009-10-30
- Keyboard problems on ubuntu
- Prepare EulerGUI 1.4
- Grunt at subversive
- Ubuntu 9.10
2009-10-29
- Solved strange random errors in Maven tests for EulerGUI
- Pause in Drools-Joafip-EulerGUI integration
2009-10-25
- Hyperlinks: ubiquity-xforms, Nepomuk, transmodel.org, ...
2009-10-24
- Create a branch in EulerGUI for joafip + Drools
2009-10-23
- Adapt a persistent storage to Drools ?
2009-10-22
- EulerGUI : on semanticweb.org; group in French
- EulerGUI
2009-10-21
- Trying Virtuoso RDF store, Open Source version
- Complete Kubuntu to Ubuntu migration
2009-10-20
- Biological Collections software
- Generating XForms with EulerGUI
- Capturing desktop session for demos
- Looking for tools for doing JavaScript animations
- Troubles with webcam Philips SPC210C on Ubuntu x86_64
2009-10-19
- EulerGUI Jena output stack trace
- Generating XForms with EulerGUI
- Links: TAP, etc
- Solving problems with 2 screens on Ubuntu
2009-10-16
- EulerGUI features
2009-10-15
- Déductions & XForms
2009-10-14
- The tests for N3 ParserJ4 from rdflib
- EulerGUI feature: read plain XML from URL
2009-10-13
- EulerGUI features
  - Detect RDF, OWL and plain XML
  - XML import
2009-10-12
- Déductions GUI rules: generate XForms
2009-10-11
- EulerGUI new feature: XML Export
- EulerGUI is a hybrid creature
2009-10-09
- EulerGUI new feature: XML Export
2009-10-08
- Xorion : a CMS with XRX architecture
- Communication
2009-10-07
2009-10-06
- Trying dbpedia
2009-10-05
- Trying Sesame 2.3-pr1 SPARQL server
- Trying Joseki SPARQL server
- EulerGUI new feature: calling SPARQL end point
  - Exploring SPARQL public endpoints
2009-10-05
2009-10-01
- EulerGUI new feature: calling SPARQL end point
- Maven + Pellet
2009-09-30
- EulerGUI new features
2009-09-29
- Trying DL queries in Protégé 4
- Trying Pellet SPARQL-DL queries with Pellet
- EulerGUI : change the description on main page
2009-09-28
- XML & XSD import : update
- EulerGUI: specifications
2009-09-27
- EulerGUI: Reading plain XML files or XML Schema
  - Mail to Roberto G.
2009-09-25
- Links: Java sources
- Maven ecosystem
- cimtool.org
- add Maven config. in ReDeFer
- Crash with Jena 2.6.0 and HSQLDB
2009-09-24
2009-09-23
- Déductions: plain XML as RDF
  - Testing Redefer plain XML to RDF conversion
    - Problem with attribute namespace
2009-09-22
- Missing feature in Gloze: enumerations
2009-09-21
- triplr
- EulerGUI instantiation: correction
- Déductions GUI rules: migration after package change in EulerGUI
2009-09-20
- EulerGUI : refactoring : moving Project and N3Source in a separate package
- Java instantiation in EulerGUI
- Trying rhizomik / Redefer project
2009-09-19
- Déductions GUI rules: migration to generic GUI platform
2009-09-18
- Déductions: plain XML as RDF
  - Trying to debug the output of Gloze
  - Trying rhizomik / Redefer project
- Examine XForms for integration with Déduction GUI rules
2009-09-17
- Déductions GUI rules: migration to generic platform
- Links: Dbin
- Tried Gloze on a large XSD
2009-09-16
- Installing Sonar
2009-09-15
- Exploring Jena
  - Exploring Jena code base
  - Trying jena eyeball
- Amaya : features wished
- Fuxi
  - Fuxi and N3 syntax
  - Fuxi has moved
2009-09-14
- Trying SGML/XML with Yap
- sourceforge download stats KO
- Trouble with Philips Webcam SPC210NC
- Links:
2009-09-04
- My program for "vacancies"
- Announces for EulerGUI 1.2.1
  - Announce in the Protégé list
- Announce for the Drools list
- Announce for the PlanetMDE list
- Announce for the news AI list
- Announce for news:fr.comp.lang.java
- Announce for LinuxFr.org
2009-09-03
- Déductions generated applications : test and debug
  - Test cases for Déduction from Web
2009-09-02
- How to find any eclipse source file
- Converting UML 2 metamodel into OWL
2009-09-01
2009-08-31
- Translating Java class into OWL
- Convert UML2 metamodel into OWL
2009-08-30
2009-08-28
- OWL 2 with form generator: TODO
- Google Summer of Code 2009
- Prolog
  - Prolog design patterns
2009-08-26
- Pyke, Logic Programming in Python
- Spell check in different language in Amaya
2009-08-25
- Links: fresnel, etc
- Releasing EulerGUI with Maven
2009-08-24
- Conzilla - A Concept Browser
- Links: graph transforms, etc
- EulerGUI mailing list management at sf.net
- N3 refactoring: extract part of antecedent
- Déductions generated applications : choice of a subclass
- Problem with Drools recompilation
2009-08-21
- Déductions generated applications : choice of a subclass
2009-08-20
- Rules about rules with Euler / Eye
- Déductions generated applications : choice of a subclass
2009-08-19
- Déductions generated applications : choice of a subclass
- Memory analyser
- Maven eclipse integration
- N3 syntax for jEdit
2009-08-15
- RDFS to OWL conversion - food example
2009-08-14
- Déductions generated applications
2009-08-13
2009-08-08
- Article for OWLED 2009
- FOAF: when owl:Thing is specified as a placeholder
2009-08-05
- Déductions generated applications
2009-08-04
- Déductions generated applications
2009-08-03
- Déductions generated applications
2009-08-02
- Generate readable diagrams
- Déductions generated applications
2009-07-31
- Déductions generated applications
2009-07-30
- Déductions generated applications
2009-07-29
2009-07-28
2009-07-27
- Déductions generated applications
2009-07-26
2009-07-24
2009-07-23
- Misc. link on RDF, OWL
- XForms
2009-07-21
- Déductions generated applications
  - Idea: JCHR
  - Update
2009-07-20
- Idea : intelligent window manager
- Add Fuxi inference engine in EulerGUI
- Trying CHR as a production rule system
2009-07-17
2009-07-16
- TODO
- Déduction generated application new architecture
2009-07-15
- OWLED 2009 - OWL: Experiences and Directions
2009-07-14
- Links : B method, etc
2009-07-06
- Generating XML from RDF
- RETE engines in Phyton
- Drools N3 engine: Duplicated triples
2009-07-04
- CALO, Cognitive Assistant that Learns and Organizes
- The tones project
- OWL 2
2009-07-03
- Drools N3 engine
  - What
  - How
2009-07-01
2009-06-29
- The Sourcerer project
- Inference engine with database access
2009-06-28
- Explanations on N3 + Java in Déductions generated applications
- Building Protégé from sources
- Proposals for the Protégé team
  - Use RDF/turtle format instead of XML or Java properties files
  - Leverage on semantics and rules for Java insfrastructure
- Compare Euler Eye engine and Drools N3
2009-06-27
- TODO Déduction form generation new architecture
- Trying Sindice
- Back from Protégé conference 2009 in Amsterdam
2009-06-21
- UML Textual (UMLT)
2009-06-20
- Tried opencalais
- Open Mind Common Sense
2009-06-18
- Readings about logic
2009-06-17
- Déductions app. generator: calling user Java code
2009-06-16
- Ontology links
2009-06-13
2009-06-12
2009-06-10
- Twitter message
2009-06-08
- Trying Protégé 4.0.111 with an N3 file
- Déductions: mixing languages in rules
- e-commerce demo.
- My first FOAF description
2009-06-04
- Déductions: testing new generated GUI
2009-06-02
- Sound is silent on Ubuntu
2009-06-01
- CouchDB
- Idea: Using Déduction Forms for creating a Déduction application
- Software categories
2009-05-27
- Testing APE (ACE, Attempto)
  - With the command line
  - With the Web client
- Oslo
- Conferences
- Making invitations for his seminars in Paris
- Changing status on LinkedIn
2009-05-24
- Misc. links
- Guice dependency injection
- Updated EulerGUI summary and categories
- Compare modelization OWL UML
2009-05-23
- Tidying HTML
2009-05-21
- Uploading stuff to Maven central repository
2009-05-20
- Setting time on Ubuntu wth ntpdate
- Hyperlinks towards EulerGUI
- Misc. links
2009-05-19
- RDF databases and applications built on them
- Groupware
- Bug tracking systems
2009-05-14
2009-05-12
- Input file for EulerGUI
- Déductions: reflexions on multi-platform generation
2009-05-08
- Tasks
2009-05-05
- Challenges
2009-05-04
2009-05-03
- Deductions
  - Command-line User Interface
  - Design GUI - KB architecture
2009-05-02
2009-04-30
2009-04-29
2009-04-28
- Problems with Maven release:perform
2009-04-27
- Using Abbot on EulerGUI
2009-04-26
- UMLT == UML Textual
- EulerGUI demo. : Facebook rules
- EulerGUI release
2009-04-25
- Sesame first steps
- N3 rule to change a literal type ?
- aaai08 proceedings available !
- Trying OpenCyc OWL Web interface
2009-04-09
- Déductions
  - New use case
  - Developers' training
2009-04-08
- Trying Twinkle SPARQL GUI on DBPedia
- morla rdf
2009-04-07
- Integrate ACE (Attempto) in EulerGUI
- Slides presented at TAO Industrial Workshop
2009-04-04
- Discussions on OO (Object Orientation) versus DL (Description Logics)
- Added new mailing list for Déductions
2009-04-01
- Connect on SourceForge with ssh
2009-03-29
- Testing RDFS to OWL conversion
2009-03-27
- Public repository for reusable models ?
2009-03-26
- Maven extra sources and eclipse
2009-03-24
- Searching for a generic XMI parser
2009-03-23
- Idea for rules applied to UML
- Problems with Maven repo. for eclipse
- XMI to N3 translator
  - generic emf factory for any URI
- UML input integration in EulerGUI
2009-03-21
- XMI to N3 Converter
- EulerGUI Demo.
- Reading and thinking on EMF
2009-03-20
2009-03-19
- Déduction architecture
- Design Java - N3 mixin
- Tried Netbeans profiler on Parser4J ( NetBeans eclipse comparison BTW)
2009-03-18
- Mixed paradigm architecture for Déductions updated
- Building IsaViz from sources (RDF editor)
2009-03-16
- Installing eclipse 3.5 M6
2009-03-16
- Development plan and status for Déductions updated
2009-03-15
- Lazy load of jars - Maven ?
2009-03-14
- Déductions : debug inheritance in forms
2009-03-13
2009-03-12
- Ideas for Euler GUI
- Misc. : pres. on rule-based programming, Science Commons, etc
2009-03-11
- Tests with Drools engine + new N3 parser
2009-03-10
- Déductions generated GUI
- Integrating new N3 parser with Drools engine
  - Integrity Constraints
2009-03-09
- Discussion with Pieter Van Gorp - grabats 2009
2009-03-06
- Ideas : JoaFip, EMF, JSR94
- LinkedIn current status
2009-03-05
- Challenges of Deduction project: calling Java
- EulerGUI : Drools engine - findall
2009-03-04
2009-03-02
- EulerGUI Generated application
- Test VNC - serveur Linux, client Windows UltraVNC
2009-03-01
- Deductions: to do before wednesday demo.
- EulerGUI: Current data binding partipating classes
2009-02-27
- Deduction runtime: storage
2009-02-26
2009-02-25
2009-02-24
- Proposal for a vocabulary for RDF projects for development tools
- Announce of EulerGUI 1.0
  - Announce on news:comp.lang.java.announce
  - Misc.
- Ideas
2009-02-23
- Problems with Maven release:prepare solved
2009-02-22
- Problems with maven-assembly-plugin + jar-with-dependencies
2009-02-19
2009-02-18
- Drools bug ?
2009-02-17
- On the path to a Drools bug
2009-02-12
- Problem solved with Drools
2009-02-11
- Testing APE
- Debug with Drools
- Drools engine for N3 logic: Drools interrogations
2009-02-08
- Drools engine for N3 logic
2009-02-07
- Workshop on Controlled Natural Language (CNL) 2009
- EulerGUI project
  - Drools engine for N3 logic
  - Use N3 format ( TODO )
2009-02-06
- SQL humour
- Drools engine for N3 logic applied in "Déductions" GUI generation
2009-02-05
- Drools
2009-02-04
- Prolog SWI - Yap compatibility
2009-02-03
- Debugging my rules for GUI construction
2009-02-02
- Debugging my rules for GUI construction
  - Tests
  - Ideas and remedies
2009-02-01
- Test OWL Inconsistent ontology
- Debugging my rules for GUI construction
2009-01-31
- Tidy EulerGUI
2009-01-29
- Déductions - GUI generation : popup frame for object properties
- Trying a large OWL with Déductions GUI generating rules
2009-01-28
- Déductions - GUI generation : data binding
- Euler GUI
- Links: AI
2009-01-25
- N3 block comment
2009-01-23
- Déductions - development plan
- Experimenting with e:findall in Euler
- Open an N3 ontology in Protégé 4
2009-01-22
- Euler suggestions
2009-01-21
- Customer with existing Access application
2009-01-20
- Déductions - GUI generation : work on subclasses
- EulerGUI - documentation
- Errors in Euler / EYE (new stand-alone Yap engine)
2009-01-19
- EulerGUI - integrated new Prolog self sufficient engine
- Déductions - GUI generation : work on subclasses
2009-01-18
- Déductions - GUI generation : implement Property details
2009-01-16
- Déductions - GUI generation : purchaseorder.owl
- JDK7, FAM
- Misc. confguration: NTP, SMTP
- The Property - Class graph
2009-01-15
- Déductions - GUI generation : splitting rules
2009-01-14
- Déductions - GUI generation driven by logic rules
2009-01-13
- Déductions - graph transform driven by logic rules
- Using RDF ID when no rdf:label - Generated GUI - déductions
2009-01-11
- Generated GUI - déductions
- EulerGUI - TODO - new window
2009-01-08
- Deductions architecture
2009-01-05
- Looking at RDF Java API : jrdf
2008-12-24
- Deductions architecture : generate other GUI's than Swing
2008-11-11
- Translating N3 into Drools language
2008-11-06
- Dialog on the Drools IRC
2008-11-02
- Deductions project
- Advanced JavaScript within JDK
2008-10-19
- Deductions project
2008-10-07
- Deductions project
  - Reading N3 logics specification
  - Write non-regression tests
- Componentization of software file search feature
- Google Chrome
2008-08-24
- Trying CWM proof checker
2008-07-07
- Euler N3 reasoner
- SWI Prolog
2008-05-19
2008-05-08
2008-05-04
- Building Protégé 4 from SVN
2008-05-01
- AI links
- Installing Coq on Fedora
- Workflow for a DL + Prolog framework: trial of Euler
- Questions on Euler
  - Notes
  - Questions
- The good servant: example
2008-04-20
- Trying tabled Prolog: XSB
2008-03-29
- Trying Sigma Kif Web application
- Programming advice: let the semantic be explicit
- Representing weight in FOL
- Upper ontologies
  - SUMO
2008-03-22
2008-03-19
- OCL and First Order Logic
- Prover9 : how to recover the answers
2008-03-09
- CHR
- Blog versus HTML site
2008-03-02
- CHR, SQL+ XSB
2008-02-17
- CHR
2008-02-03
2008-01-28
- Intentional programming; describing libraries formally in term of functionalities
2008-01-27
- My gnuplot tricks

	Language	Comment
Trac	Python	looks good, many features and users
codendi	?	supposedly Open Source, but looks like a closedsource software
redmine	Ruby	Written using Ruby on Rails framework; no task management and doc. management
mantis	PHP	bugs only, but adding plugins is possible
bugzilla	PHP	extensive search capabilities; not very ergonomic, no tasks
flyspray	PHP	apparently no documentation
scarab	Java	from the same forge as ArgoUML

2010-02-07

List of Projects That Use Protege

Gloze problems

The gloze.base parameter

Problem with xml:lang in the XML Schema

Problem with the XHTML Schema and string content

A bad warning message

Gloze within EulerGUI

The XML rood node is an RDF blank node

2010-02-05

RDFa

2010-02-03

To read

2010-02-02

Haskell

Recent EulerGUI commits

2010-02-01

Recent EulerGUI commits

A multi-valued property that is a LinkedHashMap: question on ##swing

2010-01-30

Specification of EulerGUI Server : parametrized queries

2010-01-27

Installed Thea

Running FuXi from repository: trouble

Links

2010-01-25

Readings: ACE View, Jena 2 Inference support

2010-01-22

Misc. links on OWL reasoners

(Re)Tried Open Mind / Common sense

2010-01-21

Recent EulerGUI commits

2010-01-20

Announces EulerGUI 1.4.1

EulerGUI bug: creating a new SPARQL query

2010-01-19

(re)trying Euler server

2010-01-18

2010-01-13

Trying Joseki from sources

A server that wraps EulerGUI

Updated Fuxi

Subtleties of regular expressions

Drools N3 engine: factorial test

2010-01-11

RDF servers

2010-01-10

2010-01-08

EulerGUI 1.4 Announce for English language followers

EulerGUI 1.4 Announce on Protégé list

Drools N3 engine: built-ins

2010-01-06

Trying TDB Jena

Intelligent test generation

ordered lists in OWL

Links: Non-monotonic_logic, Closed_world_assumption, etc

2010-01-05

Attempto (ACE)

2010-01-04

Releasing EulerGUI 1.4

MVEL

RIF (W3C Rule Interchange Format)

Attempto (ACE): tried some examples

Tests with empathy and Google Talk

introductory material on the semantic web

neo4j

2010-01-03

Checking EulerGUI dependencies

RuleML

RuleML and ACE

Attempto and N3

2009-12-24

Consolidating EulerGUI

2009-12-23

Gloze

2009-12-22

Gloze use cases

2009-12-21

Gloze

2009-12-20

The `gloze.base` parameter