From: cbmvax!uunet!m2xenix!onion!tessi!loop!dont (Don Taylor)
Message-Id: <9105220547.AA21590@loop.UUCP>
Date: Tue, 21 May 91 22:47:48 PDT
To: lojban-list@snark.thyrsus.com
Cc: dont
Subject: It looks like my request was misunderstood
Status: RO


(I hope there are not a dozen copies of this on the way
 to the list, the mailer died over the weekend but not
 before dozens of attempts to send each message)

From the replies it really looks like I did not make myself clear in my
requests.  I first presumed there was a misunderstanding between Mr. Bowen
and myself.  Now the misunderstanding appears to be more widespread.

First off, I have no intention of modifying the language, lojban.  My intention
was to push what appeared to be non-formal prose about elidables into formal
notation in the ebnf and watch what happened to the formal grammar.

Secondly, using yacc and understanding bnf and nit picking about formal
syntax is not an issue for me.  That has been my trade.  Even if it were a
problem, there are any number of books to teach people how to use yacc.
Having enough precise information about the language to capture the language
in the formal notation is my problem.  It is unclear to me if there is enough
information for someone to always decide exactly when a `cu', or some other
elidable, is or is not required.  I don't think I know enough to answer this
question yet, if I limit myself to strictly relying on formal notation.  If I
don't limit myself to formal notation then it would seem I must have total
knowledge of the language to make such decisions.

Third, to be precise, there are any number of formal grammars that all specify
the same language.  But the ebnf appeared, lacking formal notation specifying
the elidables, to define a language that was either a subset or a superset of
the language lojban.  I was looking to see if it were possible to construct a
grammar that came closer to capturing exactly the language lojban, as defined.

If elidables were taken as always required in the formal notation then
presumably the notation defined a set of legal sentences in the language but
presumably there were legal sentences in the language that were were not
captured in the formal notation.

If the elidables were taken as always optional in the formal notation then
presumably the notation defined a set of sentences some of which were outside
the legal language lojban.

I do not want ANY of this to be taken as criticism of Mr. Cowen or his efforts
or his notation or of anyone else working on the language or the grammar.  I
certainly didn't think anything was put in the grammar for fun.  I had a very
narrow goal that I was pursuing that may have little to do with other work
being done.

I hope this expanded explanation has better described my intentions.

Lastly, I would like to thank Mr. Cowen for generously
answering a number of my dumb questions by mail.

Thanks
Don Taylor
503-235-6853
loop!dont@tessi.UUCP
dont@loop.UUCP
tessi!loop!dont@nosun.west.sun.com