From - Mon Feb 26 10:37:00 1996 Received: from access4.digex.net (ql/6O0AY1b.Cw@access4.digex.net [205.197.245.195]) by locke.ccil.org (8.6.9/8.6.10) with ESMTP id HAA22066 for ; Sun, 25 Feb 1996 07:28:40 -0500 Received: (from lojbab@localhost) by access4.digex.net (8.6.12/8.6.12) id GAA18302 ; for ; Sun, 25 Feb 1996 06:53:15 -0500 Date: Sun, 25 Feb 1996 06:53:15 -0500 From: Logical Language Group Message-Id: <199602251153.GAA18302@access4.digex.net> To: lojbab@access.digex.net, veion@xiron.pc.helsinki.fi Subject: Re: resolution of open Lojban issues - limited distributiuon message Cc: cowan@ccil.org, pcliffje@crl.com X-Mozilla-Status: 0011 Content-Length: 1835 >About termsets: > > Termsets can be eliminated from the YACC grammar if we adopt > some kind of sumti glue along the lines of my intmed grammar. > The glue is marked with bo in the following schema: > > nu'i sumti sumti nu'u .e sumti sumti nu'u > => sumti bo sumti .e sumti bo sumti > > nu'i ge sumti sumti nu'u gi sumti sumti nu'u > => ge sumti bo sumti gi sumti bo sumti > > The glue would carry no semantic value. It doesn't eliminate temrsets, but it might eliminate the need for the bracketing nu'i/nu'us. There is no construct that allows "term" as opposed to "sumti" inside a logical connective, and indeed GEKed sumti are down in the middle of the sumti grammar at rule 93. >>About termsets 2: >> >>BEI can be used as glue at the terms level. So if we exclude termsets >>from within tanru linkargs, we get a rather regular system: termsets >>would use bei just like linkargs do. >> >>terms_80 : term_81 >> | terms_80 term_81 >> | terms_80 BEI_505 term_81 >> ; >> >> mi klama le zarci bei le briju .e le zdani bei le ckule > >It would be nice, but I think we need tanru termsets, if for no othe >r >reason than tanru being used in SEI/FIhO constructs. But then I cannot >think of too many places where you would want termsets in such things, and >you could always include them before the selbri. No strong feelings, but this really requires others besides me to think about. Reusing existing constructs is nice, but I am of course resistent to overload as in the case of Jorge's ke'a. Indeed, i still prefer adding ce'e instead of using BO for the sumti link in whatever termset grammar we end up with. I will run these by Nora, though. lojbab