Parser and Glosser Beta versions now available

[Bob LeChevalier-Logical Language Group <lojbab@lojban.org>]

I am pleased to report that we are making the Beta test versions of the
Lojban parser, and Nora's parser/glosser available to Lojban List members
for Beta testing.  Both are Windows/MS-DOS programs, and the Glosser uses
MS-DOS command files internally to the processing.

Because these are Beta versions and subject to change (or recall) at any
time, we are NOT linking them into the Lojban file server index page yet.
If you wish to put a link on your page, that is fine, but please include
warnings about their Beta status.

You may report problems with the parser to John Cowan at cowan@ccil.org.
You may report problems with the glosser to Nora LeChevalier care of me at
lojbab@lojban.org.  At this point we are promising to accumulate trouble
reports, but in particular John is not able to do much work on the parser,
and one should not expect rapid response on trouble reports.

This is the first post-baseline parser release, and hence has been much
needed.  But there is at least one known bug and one
suspected-but-not-yet-replicable bug, and there will likely be more.  The
known bug shows up because it causes the glosser to misbehave in certain
cases (see my other post for one).

Nora is actually releasing two versions of the parser/glosser at once.  One
version will insert case/tags, and the other will insert prepositions, when
labelling the places of the selbri.  You switch versions by changing the
single data file used for both versions; this is documented in the file
archive.  The default version if you unzip the archive is for preposition
labelling.

Nora is welcoming suggestions for changes in the labels for either case
tags or prepositions.  This is not something that can be tested thoroughly
without a LOT of usage.  Please let us know the sentence justifying the
change if possible, so that we can build a set of test cases.

The parser/glosser includes a copy of the parser and uses the same version
of the parser being released, so if there is a change in the parser, you
will want to copy it into the glosser directory.

Because of the idiosyncracies of the parser/glosser interaction with
MS-DOS, Windows-based execution of the program (including file dragging to
an icon) can have file-location problems.  It is best to have all files
including any text file to be parsered/glossed all in the same directory as
the program.

The files are located at

http://www.lojban.org/files/PARSER.ZIP   (42K)
and
http://www.lojban.org/files/glosser3.ZIP    (227K)

Case seems to be significant in the filenames.

The parser/glosser is expecting grammatical text - parser errors will lead
to rather poorly documented error reports.  If you want to gloss a Lojban
text file that may be ungrammatical (it will of course not have
prepositions or case tags), you may in MS-DOS type the following at the
command line

gloss3 i=inputfile [o=outfile]

where the output file parameter is optional and text will go to the screen
if omitted.

The glosser program is not especially fast on older and slower machines -
Nora uses a 486/100 for development.

Enjoy.  Comments are of course welcome.

lojbab

----
lojbab                     ***NOTE NEW ADDRESS***           lojbab@lojban.org
Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA               703-385-0273
Artificial language Loglan/Lojban: 
  see Lojban WWW Server: href=" http://xiron.pc.helsinki.fi/lojban/ "
  Order _The Complete Lojban Language_ - see our Web pages or ask me.