Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-05-25

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
03:40 ronsavage joined #marpa
04:23 jeffreykegler https://github.com/jeffreykegler/kollos/blob/master/components/main/kollos/location.lua
04:23 jeffreykegler rns: I copied your location object prototype code over to kollos and hacked it up.
04:25 jeffreykegler Right now what it does is kind of pointless -- you can give it a fixed string and get it back.
04:25 jeffreykegler Sort of a write, then read, only more expensive.
04:25 jeffreykegler :-)
04:26 jeffreykegler But I think my code may give you an idea of where I am going with this "fixed string in the prototype" idea, and how it works.
04:27 jeffreykegler The object created can now be used as a basis for new objects which are pointers into the string, or substrings of it, wherever references to the input are needed.
05:06 rns jeffreykegler: Great, I'll look into it.
05:09 rns left #marpa
05:48 ronsavage joined #marpa
09:40 lwa joined #marpa
10:11 ronsavage joined #marpa
11:05 shadowpaste joined #marpa
15:07 koo6 joined #marpa
15:50 ceridwen jeffreykegler, Do you mind if I cite you in my write-up of some these questions?
16:24 ceridwen jeffreykegler, http://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1176&context=cstech
16:26 ceridwen This paper proves that LL-regular is a proper subset of LR-regular by exhibiting a language that's LL-regular but not deterministic and one that's deterministic but not LL-regular, then noting that LR-regular contains the deterministic languages.
16:45 jeffreykegler joined #marpa
16:45 jeffreykegler ceridwen: I'd be flattered to be cited
17:31 jeffreykegler btw, I am considering the idea of adding actions to the L0 rules in Kollos.
17:32 jeffreykegler I'd been reluctant because it totally blurs the conceptual difference between lexical and structural.
17:33 jeffreykegler And overuse of it will be a great temptation for those who don't have the distinction clear.
17:33 jeffreykegler But it will allow some useful hacks.
17:34 jeffreykegler And it'll fit into the way I am developing Kollos -- I am writing the lexical portion first and if it handles semantics, I can basically have all major features tested by the time I get to adding the G1 (structural) grammars.
17:36 jeffreykegler It'll also allow another kind of scannerless parsing -- though since it actually does everything into the scanner, it's really more like parserless parsing.
17:37 jeffreykegler Perhaps we should rechristen "scannerless parsing" to "seamless parsing"
20:42 ceridwen I've always been a fan of one-stage parsing, but scannerless parsing is certainly not my favorite term, especially since "lexer" seems to be more common than "scanner" in practice.
20:47 ceridwen Now that I have a better handle on the exact complexity differences between the various modern parsing algorithms (GLL, GLR, Earley, ALL(*), etc.) , I've recently been trying to find some hard data comparing them, but I haven't been successful in finding much.
20:50 ceridwen I'm interested in empirical evidence and specifics about how algorithms perform under different conditions, how well they specialize, etc.
22:28 Aria Added that to the collection!
22:43 ronsavage joined #marpa
22:44 jeffreykegler joined #marpa

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary