[antlr-interest] Re: Nondeterminism problem

[mzukowski at yci.com]

I can't think of any reason an LL grammar couldn't be one method per rule.
Do you have a counter example?

My understanding of LALL is that it "simply" collapses the search space for
lookahead tests at the expense of some accuracy.  I don't see any reason why
doing that would be related to the one method per rule question.  

Monty

-----Original Message-----
From: Oliver Zeigermann [mailto:oliver at zeigermann.de] 
Sent: Friday, December 19, 2003 8:48 AM
To: antlr-interest at yahoogroups.com
Subject: OT: [antlr-interest] Re: Nondeterminism problem

Terence Parr wrote:
> I think you'll find that LALL(k) done by PCCTS is a proper *superset*.  
> I don't understand "only one that does practical SLL(k) near 
> computation".  Can you rephrase?  Again, PCCTS does full LALL(k); my 
> dissertation was how approximate lookahead can be used to attenuate the 
> complexity of computation and storage of lookahead.  This includes 
> weaker parser as well as helping to build full LALL(k) parsers.

All this makes me wonder if there is any way to have a recursive descent 
compiler generator handling all LL grammars that still has a single 
method per grammar rule? If not is this one of the reasons ANTLR does 
LALL instead of LL?

Oliver



 

Yahoo! Groups Links

To visit your group on the web, go to:
 http://groups.yahoo.com/group/antlr-interest/

To unsubscribe from this group, send an email to:
 antlr-interest-unsubscribe at yahoogroups.com

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


 

Yahoo! Groups Links

To visit your group on the web, go to:
 http://groups.yahoo.com/group/antlr-interest/

To unsubscribe from this group, send an email to:
 antlr-interest-unsubscribe at yahoogroups.com

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