Protege 3.4 Beta update (build 500)

There is a new beta build available on our Web site:

http://protege.stanford.edu/download/prerelease/full/
http://protege.stanford.edu/download/prerelease/basic/

IMPORTANT - PLEASE READ: This beta build is a major change from the 
previous build (build #130) and is likely to have some initial problems 
because there is so much new functionality.  So - save your work before 
trying this build, don't use if for mission critical applications  ;) , etc.

The motivation for such major changes is that we have made three 
significant improvements that have been requested for a long time:

1. Database Inclusion - Until now, database inclusion has not been 
available in Protege.  This has led to many workarounds and 
inconveniences for users who want to work with ontologies in a database 
mode.  For users who are ready to try this functionality out, we have 
written a web page [1] describing how to use the database inclusion 
mechanism. Note that there has been an incompatible change in the 
database format. You will not be able to use existing database projects 
- you will need to use a pre-build 130 version of Protege to convert the 
database project to an owl project first and then convert it with the 
new build to a database project.

2. New Protege-OWL Parser - The previous version of the Protege-OWL 
parser had many problems that made it perform badly, use lots of memory 
and be fragile. The new parser is much more robust, uses significantly 
less memory, is much faster and has a better support for untyped resources.

3. Decoupled Namespace Management - There have been several complaints 
about the way that Protege handles namespaces.  For example, users have 
requested better URN support and a decoupling of the name of the 
xml:base, ontology name and the default namespace. In previous versions 
of Protege, the namespace code was tightly coupled with the underlying 
owl model.  This led to many quirks in the way that Protege behaved.  
Now the namespace management code is a thin layer which defines how the 
ontology is presented to the user.  This build provides better URN 
support and better xml:base, ontology name and default namespaces 
decoupling.

Known Incompatibilities

1. The database schema used by the database backend has changed in an 
incompatible way.  An existing database project will need to be 
converted to a file based project before it can be used with this 
version of Protege. This means that to use an existing database with the 
new build, you will have to first convert it to a file using the menu 
File -> Convert project to format. And then, open the converted file 
project with the new build and convert it back to a database format.

2. For OWL projects, the format of  the .pprj file has changed in an 
incompatible manner.  The new version of Protege will be able to read 
existing .pprj files but when these files are saved they will not be 
properly understood by previous versions of Protege.

3. Another important change is that the internal names of the ontology 
entities are now the fully qualified name, rather than the prefixed 
name. This may impact application developers. It is recommended to adapt 
your code to use the full name. The short name is still supported, but 
may have a performance impact. The support for short names can be 
disabled, by setting in the protege.properties file the property 
owlmodel.expand.short.name.in.methods to false. The class 
edu.stanford.smi.protegex.owl.model.NamespaceUtil provides conversion 
methods between the full and short names of ontology entities. If you 
are developing a user interface, then it is recommended that you display 
the browser text of an entity, which shows the prefixed name.

3. Although we tried to test as many plug-ins as possible, some plug-ins 
may need to be adapted to the new build. If you notice any plug-ins 
misbehaving, please report them on the protege mailing list.

4. If you have existing Collaborative Protege projects in OWL, you will 
not see the annotations attached to the ontology components anymore, 
because the name of the entities in the annotation ontology used to be 
the short name, rather than the long name. We will provide shortly a 
conversion routine that will adapt existing annotation projects to use 
the full name of entities.

Best Regards,
The Protege Team

[1] 
http://protegewiki.stanford.edu/index.php/Working_with_the_Database_Backend_in_OWL

_______________________________________________
protege-beta mailing list
[email protected]
https://mailman.stanford.edu/mailman/listinfo/protege-beta

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03