From cbmvax!uunet!mullian.ee.mu.OZ.AU!nsn Fri Apr 12 01:47:20 1991 Return-Path: From: cbmvax!uunet!mullian.ee.mu.OZ.AU!nsn Message-Id: <9104120350.3518@mullian.ee.mu.OZ.AU> To: lojban-list@snark.thyrsus.com Cc: nsn@mullian.ee.mu.oz.au Subject: Elision, and {la nitcion.} eats humble pie Organisation: Department of Electrical Engineering, University of Melbourne Smiley-Convention: %^) Date: Fri, 12 Apr 91 13:50:05 +1000 Status: RO >Yes, you were being so snappy and irritable that you completely >missed the point of my relatively pointless postscript: you >overlooked the seventh sentence. Go back and look at it again. Heh heh. Subtle of you. The sentence was which may be translated at {.i.i}, I suppose. Empty sentences are perfectly grammatical. > ^ The elidable terminators make the language unambiguous, but may often be > ^ ^^ ???????? > ^Did you mean "ambiguous"? > No he does not. Whatever do you mean, Guy? To summarise: Guy thought the statement said terminators {poi} elidable, whereas what was meant was terminators {noi} elidable. Ain't lojban cute ?%^) >(Do you remember that Saturday Night Live sketch No, actually, we don't get it in Australia %^( Stuff on YACC and context-sensitivity deleted. I hate to admit it but (hm) Guy has a point. The BNF treatment is too handwaving. I don't know what the full rigorous solution should be, and I consider it a pity that context- sensitivity is such a spanner in the works for parsing, but I would not change the BNF as it is for anything. I suppose a proper handling of terminators should be worked out as some appendix to the grammar (maybe a preprocessor- type job, though definitely not as code-dependent as in YACC). What I am certain of is that the elisions are necessary in lojban, and I don't want to see them go, and surely the right terminals can be inserted before the right lexemes (if C code can do it, so can an abstract algorithm), and is anybody gonna have a go at this? {doi lojbab. do'u} [note the terminator just then %^)]: do you have any documentation on this? Perhaps this should be covered before baselining? >I am proposing that it is better for some purposes, if feasible, to produce a >more complex grammar for lojban that would eliminate the need for >context-sensitive side-conditions that dictate when elision is permissible. Maybe, but I'd treat that as a last resort. An illegible grammar (like the YACC of yore) is not much use paedagogically (as you point out), and if people can't use it, who cares about the computers. >(Imagine a variant of lojban in which you could >declare "little words" on the fly!) > typedef at John; > I gave John the office. /* This means "I gave at the office." */ I'd rather not, actually, but I have this hunch this may in fact be possible. Oh well. Never mind me - I'm an esperantist, I don't know any better. Got a hold of the Loglan article in Scientific American, June 1960. Nothing like lojban, lemme tell ya. Bye all. co'omi'e nitcion.