Re: [lojban] Re: Official parser problem?

[Bob LeChevalier <lojbab@lojban.org>]

At 05:39 PM 3/17/04 -0800, Robin Lee Powell wrote:
> On Wed, Mar 17, 2004 at 08:34:09PM -0500, John Cowan wrote:
> > Robin Lee Powell scripsit:
> >
> > > To parse: "lo'u mi le'u si lo'u mi le'u"
> > >
> > > You do step "c" of the preprocessing
> > >
> > > LOhU any_words_697{mi} LEhU si LOhU any_words_697{mi} LEhU
> > >
> > >
> > > LOhU any_words_697{mi} LOhU any_words_697{mi} LEhU
> >
> > So far so good.  But this is ungrammatical, because between LOhU and
> > LEhU you have three tokens, and only one is permitted, which must be
> > an any_words_697.
> >
> > > Now, theoretically this is one big, long lo'u...le'u string.
> >
> > Nope.  It's OK for a "lo'u...le'u" string to contain a "lo'u" *word*,
> > but what you have here is a "lo'u" *token*.
> >
> > > But the grammar as writen will not accomodate it.
> >
> > Rightly so.
>
> .o'onaisai
>
> You're glorifying an accident of the technology that was used to make
> the grammar!

The grammar was written quite intentionally around the idiosyncrasies of 
the technology that was used to make it.  Among other things, at the time 
we were pushing the limits of what the YACC program we were using could 
handle with complexity - the software company (Abraxas) actually modified 
their PCYACC product for free for us because we were breaking it, even 
though we were a charity customer.  We made several choices because they 
*worked*, not because they were elegant.

It isn't "glorifying".  It is baselining.

> No actual Lojban speaker would go, "Gee, I seem to have a lo'u token in
> my input, not a lo'u word.  I better give up.".

At the time we taught that if you were going to use lo'u or le'u in the 
word string, that you shouldn't, and go for zoi quotes instead.  We 
attempted to make the lo'u/le'u preparser rules clean by considering how it 
could be broken, but we were not trying to maximize the capabilities of 
lo'u/le'u.

> Quite frankly, I'm stunned that you are actually supporting such an
> asinine position.

The comment is uncalled for.  We made decisions based on the priorities of 
the time, whether you consider them "asinine" by present standards or 
not.  No one proposed changes to this during the time when the grammar was 
up for modification under the baseline change proposal system, so no 
changes were made.  Our *job* was to defend the baseline (and still 
is).  You asked what the design was, and were told; calling the answer 
"asinine" doesn't make people want to answer your questions.

lojbab

--
lojbab                                             lojbab@lojban.org
Bob LeChevalier, Founder, The Logical Language Group
(Opinions are my own; I do not speak for the organization.)
Artificial language Loglan/Lojban:                 http://www.lojban.org