[antlr-interest] Re: dfa-based lexers versus top-down antlr lexers

[Jason Osgood]

Hi Terence Parr-


Sorry for the delayed reply.

> I'd like thoughts also from people with *no* experience 
> using DFA-based tools like lex/flex.  Do ANTLR 
> lexers seem "natural"?

I started with ANTLR, so don't know any better.  And the expression of
a BNF in an ANTLR grammar seems straightforward.  

The only reason I'm learning about DFA and LALR now is because the
interactive parsing algorithms are mostly built on top of lex/yacc. 
(I still don't know enough to apply the ideas to LL.)


My needs are simple.  In designing a language from scratch, I have the
luxury making it easy for both humans and tools.  In other words, when
my ANTLR grammar got weird, I changed the language to make things
simple again.  Hopefully I got the balance right.

But your language translation stuff is a whole other world.  I don't
envy you.  And while I consider "embedded" languages (e.g. SQL within
JavaScript within HTML) to be a sign of failure, it's apparently
unavoidable.  

So I encourage you to make the tools you need to make your own work
easier.  The rest of us will draft off your lead. <grin>


Cheers, Jason Osgood / Seattle WA


 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/