[comp.ai] Natural Language Software Registry

potter@csusac.ecs.csus.edu (david potter) (04/11/91)

Elizabeth Hinkelman announced:

<< following excerpts only a fraction of original entry; refer to it 
    for the full article >>

> Article 5016 of comp.ai:
> From: registry@TIRA.UCHICAGO.EDU (Computational Linguists)
> Subject: announcement of the NL Software Registry
> Date: 10 Apr 91 20:15:42 GMT
> 
> [please distribute this announcement -- EAH]
> 
> 
>                   NATURAL LANGUAGE SOFTWARE REGISTRY
> 
> 
> The Natural Language Software Registry is a catalogue of software
> implementing core natural language processing techniques, whether
> 
> This document is available on-line via anonymous ftp to tira.uchicago.edu 
> (IP 128.135.96.31), by email to registry@tira.uchicago.edu, and by 
> physical mail to the address below.  If you have developed a piece of 
   <<   deletions here, and below >>
> 
> Elizabeth Hinkelman, Director (registry@tira.uchicago.edu)
> NL Software Registry
> Center for Information and Language Studies
> 1100 East 57th Street
> Chicago, IL 60637, USA
> 

  GENERAL WARNING to folks who obtain this material via FTP....:  
the README file in the subdirectory where this Registry is stored 
says that the two ".ASCII" files, "BODY.ASCII" and "INDEX.ASCII," 
are ascii versions of the base file (which is provided in several
formats).
  In the places where I've worked, we always referred to an "ascii" 
file when we meant a printable file that could simply be spooled to
any printer; or we specified the printer for which the file had been
formatted.  These files contain a good many non-printing ASCII
characters, mostly control-H's used to accomplish underlining, but
also include what appear to be printer-specific escape sequences.
  If you obtain the ASCII version of these files, I suggest filtering
them before printing or running them through a text processor.

   I've sent E-mail to Registry Director Hinkelman, suggesting they
emend the README file describing their files, but I'm putting a 
note in here on the off-chance that it might save someone else 
having to deal with the rather amusing output I just got from our
printer.... *grin* 
 
   P.S. ... the database itself appears to be pretty danged useful;
recommended, as long as you elide every character with a value less
than 20<hex> prior to printing on a printer which handles escape 
sequences.  (removing underline characters would probably be a good
idea, too, since their printer output underlines by printing
"_<^H><character>", and after you've elided the ctrl-H's, you'll 
have a lot of embedded underlines.....)

   Regards to all, and thanks for the Registry/Index.   Looks like
a highly useful idea, and the work is appreciated.