Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-10-19

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

All times shown according to UTC.

Time Nick Message
00:16 idiosyncrat Demat!
00:17 idiosyncrat I discover that I can remove most mentions of the SLIF from the Marpa::R3 documentation -- it is now the only interface, so I can all it the "Marpa::R3 DSL".
01:56 ilbot3 joined #marpa
01:56 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Code paste/run: https://f.perlbot.pl/#marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today - Youtube channel: https://www.youtube.com/channel/UCYKVfGBtfTqbs1JdYq-dc5g
16:19 idiosyncrat joined #marpa
16:20 idiosyncrat https://www.youtube.com/watch?v=lFBEf0o-4sY
16:20 idiosyncrat "parsing with grammars shoot-out" - Lars DIECKOW
16:21 idiosyncrat A video of Lars 6-way comparison of parsers -- I'm pleased to report that Marpa won.
16:22 idiosyncrat The talk was given 10 August 2017 at the Perl Conference in Amsterdam
17:52 ilbot3 joined #marpa
17:52 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Code paste/run: https://f.perlbot.pl/#marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today - Youtube channel: https://www.youtube.com/channel/UCYKVfGBtfTqbs1JdYq-dc5g
21:50 ronsavage joined #marpa
21:57 koom joined #marpa
21:58 koom_ joined #marpa
22:13 idiosyncrat joined #marpa
22:16 idiosyncrat The other contestants in Lars 'shoot-out' were ANTLR 4, eyapp (a Perl yacc implementation), Perl 6, Pegex, Regexp::Grammars and something called hammer.
22:17 idiosyncrat The order of rules matters in hammer, which makes me guess that, like Perl 6 and Pegex, it's a PEG implementation.
22:18 idiosyncrat All in all a pretty thorough survey of the kinds of parsers still hanging on to life out there. :-)
22:25 idiosyncrat Apparently Larry was there, that sounds like him responding at https://youtu.be/lFBEf0o-4sY?t=18m40s
22:28 idiosyncrat His point, that in Perl 6 left recursion is considered a bug, is typical of the traditional state of the art in parsing -- ever since yacc, it's been considered acceptable to have a hard-to-identify boundary between acceptable grammars and unacceptable ones, to provide little or no error guidance, and to dismiss grammars that the parser won't parse as bugs and the programmers fault.
22:29 idiosyncrat That's why, if you go over the history, you see yacc abandoned, but with almost nobody complaining about it on the record -- if you couldn't make yacc work, you were saying you couldn't make the state-of-the-art tool work, which basically amounted to saying you were a weak programmer.
22:31 idiosyncrat So yacc's usage fell as it burned one programmer after another until the turning point when GCC when back to recursive descent -- but yacc is still taught in those schools which still offer parsing courses.
22:33 idiosyncrat Any, yacc created a sort of parsing version of the "Stockholm syndrome", whereby behaviors which would be completely unacceptable in any other tool are considered signs of sophistication in parsers.
22:34 idiosyncrat Combine this "blame the programmer" for crossing the boundary ethos, with a retreating boundary (PEG and recursive descent are weaker parsers than yacc was), and you end up with parsing technology being absolutely miserable, but almost nobody complaining.
22:38 idiosyncrat Lars says his problem reports on #perl6 didn't get much attention and this would be why -- LL and PEG are considered state-of-the-ever-shrinking-art, and inability to deal with their limits is not the implementor's issue -- and to be sure, it's not fixable without totally abandoning that approach to parsing.

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