Re: Machine grammar and elidables

[cbmvax!uunet!bigd.cray.com!dmb]

Don,

   You write:
>I would like to know how hard it would be to modify the ebnf grammar to
>handle some or all of the elidables explicitly.

>I suspect that either

>	all the elidables are equivalently difficult to handle AND
>	difficult to handle and the current grammar might not even
>	be organized in a way that would make this feasible to do
>or
>	some of them are easier than others and at least the easy
>	ones could be handled with small additions to the grammar.

>I am very interested in how the grammar grows as these changes are made.
>If each change doubles or quadruples the size of the grammar this tells
>me something valuable.

>I must admit that I don't have a lot of understanding about the grammar.
>Would it be feasible for a person like myself to make such changes or is
>the amount of knowledge of the language required beyond what I could do?
>I have received short prose explanations that give a general idea of when
>a word can be elided but these are not enough for me to use to turn into
>formal grammar.

>If I could not make the modifications is there someone else who would
>assist me in doing this or would do this outright?

   I was looking at this issue a few weeks ago.  I concluded that we could
eliminate "cu" with only a minor change in the grammar, without introducing
any ambiguities.  The cost would be an increase in the number of cases where
a preceeding "ku" is not elidable.  I ought to double check those conclusions
because they were done by hand.  I haven't had the time to type in the
grammar and feed it to my local YACC.  I suspect that the situation with
some of the other elidables may be similar; one elidable can be eliminated
at the cost of requiring another in more cases.  Such changes would also
eliminate one area of Lojban style, the choice of elidables to elide.

   As far as experimenting with the grammar, I'd say the key ingredients
are an understanding of formal grammars and access to tools such as YACC.
With those ingredients you can start with your changes, feed them to YACC,
and then look at the output.  Probably your changes will have introduced
some ambiguities and so you will need to look at them and rework the grammar
to eliminate them.  Keep iterating until you finally eliminate all of the
ambiguities.  You also need to keep checking that you don't lose any
expressiveness in the process.  The more you know, the easier all of this
is, but those basics aren't all that hard to acquire.

David Bowen