Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-03-26

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

All times shown according to UTC.

Time Nick Message
00:32 jeffreykegler joined #marpa
02:03 jeffreykegler joined #marpa
02:48 ilbot3 joined #marpa
02:48 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Pastebin: http://scsys.co.uk:8002/marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today
02:51 jeffreykegler joined #marpa
02:56 ronsavage Still recovering from ditching Data :: TreeDumper in favour of Data :: RenderAsTree. Now, MarpaX :: Grammar :: Parser can use Marpa's metag.bnf as the BNF, and a user's BNF as the input text, parser it, output $value as a tree of type Tree :: DAG_Node, and then compress the tree, giving a new tree, holding only the original (user's) BNF statements as nodes, and each token within each line of their BNF as a node. Phew! See https://gist.github.c
08:50 lwa joined #marpa
11:17 btyler_ joined #marpa
14:28 sirhouseboat joined #marpa
18:30 jeffreykegler http://bost.ocks.org/mike/algorithms/
18:30 jeffreykegler A web page on visualizing algorithms -- recommended by Paul Bennett
18:31 jeffreykegler I've yet to read that article carefully, but it gives me a thought.
18:31 jeffreykegler Those of you attempting to rebuild the Marpa algorithm, in another language for from scratch.
18:32 jeffreykegler Perhaps you'd like to create a visualization -- I'd think it be very helpful and it might get a lot of attention.
18:32 jeffreykegler A visualization of the Leo modification in particular I think might be very useful.
19:04 koo5 joined #marpa
19:26 RichardMichaels I am glad and grateful that you developed Marpa for Perl. Being a Perl fan, i am glad that our language has a state of the art parser. Your work on Marpa is appreciated by many.
20:02 jeffreykegler RichardMichaels: Thank you.
20:03 jeffreykegler The first years of my work on Marpa were very lonely, and the support now is a pleasant change and much appreciated.
21:08 ernimril jeffreykegler, I prefer a good whitepaper, but visualization is good. What I really would want to find is a bag of tricks, things like "a bitset can be a good representation of a ruleset, then you can find the next steps in O(1) using these operation, ... ". I find the current paper a bit too theoretical with lots of "this can be made O(1)", but no actual explanation.
21:08 ernimril jeffreykegler, reading the code is quite a lot of work, when you want to build it in another language
21:09 jeffreykegler It's not easy code, no.
21:09 jeffreykegler I
21:09 jeffreykegler I've focused on the upper layers, because if nobody want to use Marpa, nobody will care whether the code is readable or not.
21:10 ronsavage joined #marpa
21:10 jeffreykegler Right now, my plan is do Kollos, which will duplicate a lot of Libmarpa's work in the upper layers.
21:10 jeffreykegler For example, more will be done via rewriting.
21:11 jeffreykegler After that, Libmarpa will be simplified -- I hope for as much as a 50% code size reduction.
21:12 jeffreykegler Pending the simplification, I don't plan to do much about improving the current Libmarpa documentation.
21:12 ronsavage 50%! That's amazing.
21:12 jeffreykegler ronsavage: Note that while I have been good about not over-promising ...
21:13 jeffreykegler Code size reductions are something where I've always been over-optimistic.
21:13 jeffreykegler But much of the Libmarpa code is currently devoted to special cases, which can be moved elsewhere.
21:14 jeffreykegler And other code implements features like events, which are in Libmarpa because of the inefficiencies of the Perl-to-C interface.
21:15 ronsavage OK. And I do understand that the code will have to exist elsewhere anyway, but it's clear that you're still far from reducing the core code to a state perfection :-)
21:15 ronsavage 'state' => 'state of'
21:16 ronsavage Re http://irclog.perlgeek.de/m​arpa/2015-03-26#i_10346219. I like visualizing things too. That's why I wrote MarpaX :: Grammar :: Parser and *:: GraphViz2.
22:00 flaviu joined #marpa
22:47 koo5 joined #marpa

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