From lojbab@lojban.org Tue May 30 00:20:12 2000 Return-Path: Received: (qmail 4263 invoked from network); 30 May 2000 07:20:11 -0000 Received: from unknown (10.1.10.142) by m2.onelist.org with QMQP; 30 May 2000 07:20:11 -0000 Received: from unknown (HELO stmpy-4.cais.net) (205.252.14.74) by mta3 with SMTP; 30 May 2000 07:20:11 -0000 Received: from bob (31.dynamic.cais.com [207.226.56.31]) by stmpy-4.cais.net (8.10.1/8.9.3) with ESMTP id e4U7KAi78826 for ; Tue, 30 May 2000 03:20:10 -0400 (EDT) (envelope-from lojbab@lojban.org) Message-Id: <4.2.2.20000530030439.00b3bf00@127.0.0.1> X-Sender: vir1036/pop.cais.com@127.0.0.1 X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.2 Date: Tue, 30 May 2000 03:21:41 -0400 To: Lojban Listserver Subject: Re: [lojban] speech recognition In-Reply-To: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed From: "Bob LeChevalier (lojbab)" At 06:38 PM 05/29/2000 -0400, Invent Yourself wrote: >Every 4 months there is a flurry of interest concerning Lojban speech >recognition. IBM's ViaVoice gets mentioned, and folks are directed to read >the archives. Then there's complete silence on the idea until the next >spate. > >Does anything of value ever get done on this front? Not recalling the Via Voice thing, there probably won't be results of value until someone has money to throw at the problem. It might not take a lot of money as computer applications go to get meaningful progress, but it isn't a weekend hack job either. At one point, I think Mark Mandel of Dragon Systems said that for $50K we could have something comparable to the work they did on Klingon speech recognition. When we need special hardware and/or software, and/or a small but significant amount of money, I have been told that this is where we should use our status as a small technological non-profit to go up to the appropriate corporations and ask for what we need as a grant or donation (at the time, it was by a Sun employee right after Sparc stations first came out, who said that we could probably get one for free for our work if we could come up with a plausible reason why we needed one - I couldn't, unfortunately). But this whole arena is something I have no personal experience with, and thus far we have never identified any single application with specific enough needs that we would know what to ask for anyway. If someone with some interest in this area wants to do the work, can prepare a concept and proposal and wants to contact a possible corporate donor under my authorization, that would likely be the best way to proceed. But as long it remains a question that stops at "read the archives", it will remain an archives issue. (I will note that this is something that Nora is personally interested in, but she has no expertise in any technology more recent than 1988 for doing it, and of course she has little time for Lojban and a backlog of commitments - she is still trying to get LogFlash 2 done so people can be using it before LogFest. Maybe if people interested do some research this can be a topic of discussion at LogFest, since I think you've said you intend to come, xod.) lojbab -- lojbab 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: http://www.lojban.org