Message-ID: <31223588.1DE8@ccil.org> Date: Wed, 14 Feb 1996 14:18:32 -0500 From: John Cowan Organization: Lojban Central X-Mailer: Mozilla 2.0 (Win16; I) MIME-Version: 1.0 To: veion@XIRON.PC.HELSINKI.FI Subject: Re: GEN: almost-PROPOSAL: intervals References: <199602121648.LAA01061@locke.ccil.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 973 X-From-Space-Date: Wed Feb 14 14:18:32 1996 X-From-Space-Address: - la veion. cusku di'e > ERROR: the YACC modification I previously sent works at the YACC > level but isn't feasible in practice as the modification > ended up on the lexer side - I'm not yet sure whether it > can be done on the parser side, probably not without a > major modification. Regrettably, it can't. Whatever is done in the preparser rules (900-end) can't refer to things in the earlier rules, on pain of implementing the entire parser within the preparser. So "NOI sentence" within a tense is impossible, even though it YACCs, it doesn't fit the schema of the parser, which is to keep the compounded forms simple. Of these, tense (lexer_O) is already the worst offender, and further complications are truly intolerable. I am still thinking about the other possibilities. I weakly favor messing with "ve'i ", although I recognize the annoying doubled semantics of the i/a/u vowel plus what the sumti actually says.