Prolog+CG 2.0.13

[Ulrik Sandborg-Petersen-2]

Dear all,

Prolog+CG 2.0.13 has been released over at SourceForge.net:

http://prologpluscg.sf.net

This release has been in the works for quite a few years.  Sorry for the
long delay.

The Release notes, including changes over 2.0.12, are pasted in below.

Please note the support for Mac OS X.


Enjoy!


Ulrik



Release notes for Prolog+CG

Release 2.0.13
==============

This release is a bugfix release with major feature enhancements.

Changes over 2.0.12 include:

- Java 2 SE 5.0 or later is now REQUIRED.  Support for version 1.4.2
   has been dropped.

- All files (including Prolog files, ppcgapplet.xml files, and
   index.html files) are now saved as UTF-8, with a BOM (Byte Order
   Marker) at the beginning, in order to distinguish them from files
   made with older versions of Prolog+CG.  When loading a file,
   Prolog+CG now checks for a BOM at the beginning.  If such is found,
   the file is loaded as UTF-8.  If no BOM is found, then the file is
   loaded with whatever encoding is the current system's default
   encoding.  When saving, the file is, as was mentioned, saved as
   UTF-8.

   This was done so as to make Prolog+CG files portable across systems
   and platforms.  Before, a Prolog+CG program made on, e.g., Windows,
   would fail to load on, e.g., Mac OS X, if letters beyond ASCII were
   used (e.g., é, ä, ø).

   This release fixes that: Just load the file on the same platform
   with which it was originally made, save it again, and it will be
   loadable on all platforms *with PPCG 2.0.13 (or later)*.

   As a side effect, programs saved with this version will fail to load
   on Prolog+CG 2.0.12 or earlier.

- PPCG can now be run on Mac OS X as a native Mac OS X Java .app
   application.

- Prolog+CG can now be used programmatically in a thread-safe
   environment.  See the ChangeLog for 2.0.13 for how to do so.

- For some CG operations, including subsume and generalise, the
   "functional CG" restriction is no longer in effect.  That is, your
   CGs need not be "functional CGs", i.e., CGs that, for each concept,
   only have one outgoing or incoming relation of any given name.  That
   is, you can now do subsume and generalise on CGs that look like
   this:

   [Baker]-
     -uses->[Dough],
     -uses->[Salt]

   and expect to get meaningful results.

- Two bug fixes were made to the CG matching code.  See the ChangeLog
   for details.

- The debugger and the expert system are gone.  This was so as to be
   able to implement the above changes correctly.




--
Ulrik Sandborg-Petersen, PhD
Aalborg University, Denmark
http://emdros.org -- Emdros is a corpus query system
http://ulrikp.org -- Personal homepage


---------------------------------------------------------------------
To unsubscribe, e-mail: cg-unsubscribe@...
For additional commands, e-mail: cg-help@...

[David Whitten-2]

when do you expect to get debugging and the expert system to start working again?

On Mon, Oct 26, 2009 at 12:40 PM, Ulrik Sandborg-Petersen <ulrikp@...> wrote:

Dear all,

Prolog+CG 2.0.13 has been released over at SourceForge.net:

http://prologpluscg.sf.net

This release has been in the works for quite a few years.  Sorry for the long delay.

The Release notes, including changes over 2.0.12, are pasted in below.

Please note the support for Mac OS X.


Enjoy!


Ulrik



Release notes for Prolog+CG

Release 2.0.13
==============

This release is a bugfix release with major feature enhancements.

Changes over 2.0.12 include:

- Java 2 SE 5.0 or later is now REQUIRED.  Support for version 1.4.2
 has been dropped.

- All files (including Prolog files, ppcgapplet.xml files, and
 index.html files) are now saved as UTF-8, with a BOM (Byte Order
 Marker) at the beginning, in order to distinguish them from files
 made with older versions of Prolog+CG.  When loading a file,
 Prolog+CG now checks for a BOM at the beginning.  If such is found,
 the file is loaded as UTF-8.  If no BOM is found, then the file is
 loaded with whatever encoding is the current system's default
 encoding.  When saving, the file is, as was mentioned, saved as
 UTF-8.

 This was done so as to make Prolog+CG files portable across systems
 and platforms.  Before, a Prolog+CG program made on, e.g., Windows,
 would fail to load on, e.g., Mac OS X, if letters beyond ASCII were
 used (e.g., é, ä, ų).

 This release fixes that: Just load the file on the same platform
 with which it was originally made, save it again, and it will be
 loadable on all platforms *with PPCG 2.0.13 (or later)*.

 As a side effect, programs saved with this version will fail to load
 on Prolog+CG 2.0.12 or earlier.

- PPCG can now be run on Mac OS X as a native Mac OS X Java .app
 application.

- Prolog+CG can now be used programmatically in a thread-safe
 environment.  See the ChangeLog for 2.0.13 for how to do so.

- For some CG operations, including subsume and generalise, the
 "functional CG" restriction is no longer in effect.  That is, your
 CGs need not be "functional CGs", i.e., CGs that, for each concept,
 only have one outgoing or incoming relation of any given name.  That
 is, you can now do subsume and generalise on CGs that look like
 this:

 [Baker]-
   -uses->[Dough],
   -uses->[Salt]

 and expect to get meaningful results.

- Two bug fixes were made to the CG matching code.  See the ChangeLog
 for details.

- The debugger and the expert system are gone.  This was so as to be
 able to implement the above changes correctly.




--
Ulrik Sandborg-Petersen, PhD
Aalborg University, Denmark
http://emdros.org -- Emdros is a corpus query system
http://ulrikp.org -- Personal homepage


---------------------------------------------------------------------
To unsubscribe, e-mail: cg-unsubscribe@...
For additional commands, e-mail: cg-help@...

[Ulrik Sandborg-Petersen-2]

Hi David,

David Whitten wrote:
> when do you expect to get debugging and the expert system to start
> working again?

Actually, I had planned on not reinstating those features.


Do you need them?  If so, I _may_ be able to reinstate them, though I
may have to apply to my supervisor for funding/hours to do so.


Cheers,

Ulrik


---------------------------------------------------------------------
To unsubscribe, e-mail: cg-unsubscribe@...
For additional commands, e-mail: cg-help@...

[David Whitten-2]

I would be interested in knowing the cost.

In my experience, debugging is an essential tool for any system, and when it is removed, many people quit using the system and try to find another which meets their needs more.  There is a considerable difference between working on a system as a developer who understands the ins-and-outs of the code created, and working as a maintainer, who has to use all the tools of the system to begin to get an idea how it works.

On Tue, Oct 27, 2009 at 12:58 PM, Ulrik Sandborg-Petersen <ulrikp@...> wrote:

Hi David,

David Whitten wrote:

when do you expect to get debugging and the expert system to start working again?

Actually, I had planned on not reinstating those features.


Do you need them?  If so, I _may_ be able to reinstate them, though I may have to apply to my supervisor for funding/hours to do so.


Cheers,

Ulrik

---------------------------------------------------------------------
To unsubscribe, e-mail: cg-unsubscribe@...
For additional commands, e-mail: cg-help@...