X-Digest-Num: 181 Message-ID: <44114.181.1044.959273824@eGroups.com> Date: Fri, 02 Jul 1999 02:02:33 -0400 From: Bob LeChevalier-Logical Language Group Subject: Updated (FASTER!!!) Parser/Glosser version available X-Yahoo-Message-Num: 1044 Content-Length: 1639 Lines: 39 Nora and I, finally having time to look at the parser glosser this week to find out why it was running so slow, discovered some unnecessary and grossly slow code in the innermost loop of the Glosser. The program was converting all strings to upper case before doing a string compare - usually a sound safety check, but unnecessary for the strings we are comparing in this program. Removing this conversion has shortened time in the glosser to less than 1/3 of the former eon that it was taking, and Saki's The Open Window, an 8k file which is at about the upper limit for token space for the Lojban parser, takes under a minute on a 486-100 Windows machine (it was taking almost 4 minutes before the change). Other experiments to speed up the processing had only tiny effects, so we don't think much further improvement can be made, but I'm sure people will be glad for this one. The file is at http://www.lojban.org/files/glosser3.ZIP (226K) This address is NOT on the file archive roadmap. See my message of 8 May on the archive for more details on the program. This remains a Beta version - please make this clear if you link to it and/or put a copy on your site. As before, comments and bug reports are 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.