On Fri, Jul 25, 2014 at 3:06 PM, Jeffrey Kegler < [email protected]> wrote:
> Actually, the SLIF allows angle brackets, but they are optional for > non-terminals. > Yes it does! So the only thing needed is to change double quotes to single for literals and wrap literals in single quotes and that's it — https://gist.github.com/rns/17853fa8f539f4f964c9 — ah, and he uses C-style comments, while SLIF uses Perl's. :) > His analysis of the types of ambiguity is interesting. What I call > factoring he calls "associativity", and what I call "symbolic choice" (or > symch) <http://search.cpan.org/%7Ejkegl/Marpa-R2-2.087_001/pod/Glade.pod>, > he calls alternate routes (or multiple paths). My problem would be that > different factorings *also* allow alternate routes, so that I find the > linked-to page's terminology to be kind of confusing. Nonetheless he does > have the taxonomy, one which I had to discover, for doing the abstract > syntax forests. I wonder if Prof Heckendorn rediscovered it, or if it's > stated clearly in the literature somewhere and I'd just missed it. > All the more interesting that the course Prof. Heckendorn teaches <http://marvin.cs.uidaho.edu/Teaching/CS445/> seems to be, well, rather unambiguous. It was certainly implicit in the literature. "Factoring" was distinguished > as a special type of ambiguity, so by implication there are factorings and > non-factorings. > Well, nobody has had tools to try that in practice I think. > > > -- jeffrey > > > On 07/25/2014 03:31 AM, rns wrote: > > Messed up the link, sorry, this one works: > http://marvin.cs.uidaho.edu/Teaching/CS445/grammar.html > > On Friday, July 25, 2014 1:29:59 PM UTC+3, rns wrote: >> >> http://marvin.cs.uidaho.edu/Teaching/CS445/grammar.htm— that is. Very >> gentle and informal, but dense and informative, covering (E)BNF, >> precedence, associativity, ambiguity, tips for grammar writing and useful >> examples. >> >> What I find both useful and amusing: >> >> (1) the grammar syntax is nearly SLIF, sans the angle brackets; so >> >> (1a) examples can be fed to Marpa::R2::Scanless::G almost as they are; >> but >> >> (1b) 'grammar' is not described as a thing that can be used easily to >> 'parse', e.g.: >> >> A grammar is used to specify the syntax of a language. It answers the >>> question: What sentences are in the language and what are not? >> >> >> *Parse *is to show how a sentence could be built from a grammar >> >> >> A sure sign of pre-Marpa days when grammar were written mostly as a >> specification for humans, as was frequently discussed in this group. >> > -- > You received this message because you are subscribed to the Google Groups > "marpa parser" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > > > -- > You received this message because you are subscribed to the Google Groups > "marpa parser" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "marpa parser" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
