Re: LALR

On Mon, 24 Jul 2000, Chip Salzenberg wrote:

> According to Dan Sugalski:
> > Embeddability's also orthogonal to any syntax issues.
> 
> Well, almost.  A simpler language will, in general, have a smaller and
> faster parser, thus causing a smaller impact.  But I wouldn't sweat
> this as a practical issue.  Considering Moore's Law and the planned
> release date of Perl 6, I don't think we should obsess over the size
> of perly.o.

Heck, I'm not sweating the size of perl6shr.exe in toto, let alone
perly.obj's size. Though I am using embedding in the "I embedded perl in
my circuit layout app" sense, rather than "I embedded perl and X in this
teddy bear" sense. The latter does have size issues, but I think it's
reasonable to have a backend that takes in an optimized bytecode stream
and executes it. (My intuition is that parsing sucks up more dynamic
memory than anything else, so code size isn't as big a problem. Of course,
like most intuited things, I expect I'm wildly wrong...)

					Dan