Projects

[pycyn@aol.com]

Herewith a list of projects that some has mentioned a need for but which 
Lojban Central (i.e. Lojbab, Nora, and Cowan, at most) do not have 
time-energy-resources for)

fu'ivla dictionary
A list of fu'ivla isolated as of February, 2000 at least is in the draft 
dictionary/working 
directory. The words are undefined, so need translations. As a Central 
project this ranks well behind giving meanings and place structures to lujvo 
and cmavo compounds.

Lojban word processor
The original suggestion was for a Wordpad-style word processor with a lb 
spell checker that would also provide place structure for at least gismu on 
demand and would check and analyse lujvo. A checker for correctness of cmene 
also fits this early pattern. A second step was to build in the parser to 
check sentences for grammaticality and structure, and, as a side benefit, 
check for right structure closers. After an enthusiatic start the discussion 
moved to questions about how to do this: write from scratch in Java or build 
onto Emacs or Framemaker or.... This led to two other lines of development: 
further flourishes (tab completions, LaTex markups, conversion to HTML and 
the like, ...) and typography (lb in Tengwar or monospace or Hangul or...). 
These latter came to dominate the discussion. After a few people reported 
initial modules, nothing more seems to have happened. Is there a functioning 
significant part of an lb word processor out there somewhere?

speech recognition programs
Given Lojban's purported isomorphism of sound to writing and unique 
decomposition, this seems a natural. But the equipment is said to be pricey 
and expertise in its use limited. Any takers? There are several sites on 
the web which contain Lojban sounds, of various dialects, to use as guides 
and tests.


MEX testing
The MEX grammar of lojban is said to enable a speaker to read any 
mathematical formula in an unambiguous way and more simply and directly than 
the corresponding ordinary lojban or English. Is this true? This requires 
someone who knows a goodly swath of mathematics to master MEX and put it to 
the test. Lojbab has suggested starting with early childhood learning 
materials and working up step by step to see how well the system works at 
each level and to develop a facility at using it. 


Logflash
This still works well on most systems that can fake MSDOS, but several people 
have set out either to translate it from Turbo Pascal or to recreate it in 
some UNIX friendly format or a roughly universal one. It also needs, even in 
its present form, to be extended for teaching rafsi and lujvo construction. 
No one has reported success or serious progress on any of these projects. 
The most recent reports are of a started java version by jewel@pixie.co.za 
(John Leuner) and an ongoing python version by taral@taral.net (Taral) 
og-@troglodyte.org is working on what he calls huweel.