Re: LANG: VRML 1.x Binary Format Proposal

Mark Waks ([email protected])
Thu, 1 Jun 95 10:58:27 EDT


Brian points out:
>What no one has talked about on the list yet is whether a binary file
>format could provide a benefit other than compression - parsing speed.
>Compiled C programs are very rarely smaller than their source, yet no one
>disputes than an object code interpreter is faster than a source code
>interpreter :) Does anyone think there might be a gain that could be
>made in this area from a binary format?

Some, definitely. But probably not a lot. One of the advantages of
Inventor ASCII format (and one of the highest reasons I was arguing
strongly for it) is that it is *very* easy to parse. It's almost
completely unambiguous syntactically (maybe completely unambiguous),
and not even terribly ambiguous lexically. It's sufficiently
lightweight to parse that I suspect parse time isn't more than a
*tiny* fraction of rendering time, much less transmission time...

(Of course, I'd be interested in seeing any evidence to the contrary.
But I suspect I'm right here.)

By contrast, C is far from simple to parse, with lots of switchbacks
that need to be decided on the basis of context and what comes later.
In VRML, you can probably always decide what data structure you're
building on the basis of the current token, and tokens are never (?)
ambiguous in type. The only programming language I know of that's
that simple is Lisp...

-- Justin
Who is involved in the project to write the
world's fastest Ada parser, and is *quite*
conscious of how hard it can be...

Random Quote du Jour:

"While I'd hardly say I'm morally bankrupt, you could probably make
a good case for me being in moral receivership."
-- Richard Sexton