Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-08-25

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

All times shown according to UTC.

Time Nick Message
02:13 idiosyncrat_ Anyway, resuming a bit ...
02:14 idiosyncrat_ Larry Wall's Perl is the result of the "Let's just hack away at the problem, and never mind a lot of preliminary theory and math"
02:15 idiosyncrat_ I believe Larry was influenced in this by noting that is the way that human languages evolve.
02:15 idiosyncrat_ The result in Perl 5, was a stunning success, one that changed computer languages and usage forever.
02:17 idiosyncrat_ If you entered the industry after Perl 5, that may not be apparent, but that's because everybody after Perl built on Larry's ideas.
02:17 idiosyncrat_ So the Perl community has an ethos of "Never mind all the theorizing, hack" ...
02:19 idiosyncrat_ And that is the opposite of the way Marpa was created -- In doing Marpa/Kollos, I do theory first, and only once I've nailed down my theory do I hack.
02:20 idiosyncrat_ So that's a reason that the Perl world is slow to warm up to something like Marpa.
02:20 idiosyncrat_ Add to this, that, when it comes to parsing, the record of the CS theorists has been a *disaster*.
02:21 idiosyncrat_ Top-down parsing was invented by hackers, but then the theorists invented LALR/yacc, and forced it into almost total intellectual dominance and widespread use for 4 decades.
02:22 idiosyncrat_ At which point everybody admitted, that LALR actually did not work very well, and went back to top-down ...
02:23 idiosyncrat_ basically rolling back the clock 40 years and tossing 2 generations of parsing textbooks into the trash.
02:24 idiosyncrat_ So those hackers who say that parsing theory is worthless, and that to write a parser you just hack ...
02:24 idiosyncrat_ well, I think they are wrong ...
02:24 idiosyncrat_ but they sure have a lot of history on their side.
03:02 mauke_ joined #marpa
03:08 CQ_ joined #marpa
04:02 Aria Yup.
04:40 CQ idiosyncrat_: I thought a little more about the simple demos... something like Date::Extract would be interesting to write in Marpa... see the build_regexp routine in https://metacpan.org/source/SHARYANTO/Alt-Date​-Extract-SHARYANTO-0.05.01/lib/Date/Extract.pm
04:42 CQ maybe the way to convince Perlers is to have a more Do::Something::Marpa modules
04:43 CQ they make good learn ing examples as well
05:15 CQ afk for a while
05:33 pczarn joined #marpa
05:34 ronsavage CQ: A much more meaty example is the Gedcom definition. Google for Ged551-5.pdf. It may be a zip file Ged551-5.zip or gedcom55.zip.
05:38 ronsavage Nevertheless, build_regex is an interesting bit of code.
06:18 CQ2 joined #marpa
06:30 pczarn joined #marpa
07:17 idiosyncrat_ Re possible demos/projects, I will refrain from saying a lot
07:18 idiosyncrat_ At one time, I thought that coming up with ideas of possible projects was a useful thing to do, and it was on Ron's website.
07:19 idiosyncrat_ But folks came up with their own ideas, and my list was ignored, which was probably fortunate.
07:21 idiosyncrat_ A short while ago, I asked Ron about it, and we agreed the list was hopelessly outmoded now, and deleted it.
07:21 idiosyncrat_ Good night!
07:29 koo7 joined #marpa
07:59 koo7 joined #marpa
13:23 koo7 joined #marpa
13:45 pczarn joined #marpa
14:18 pczarn joined #marpa
15:16 koo7 joined #marpa
15:39 pczarn joined #marpa
16:11 idiosyncrat_ joined #marpa
17:23 lwa joined #marpa
17:35 purmou joined #marpa
18:22 djns joined #marpa
19:24 mvuets joined #marpa
19:57 lwa I have given the naming of the upcoming MarpaX::Grammar::Preprocessor some thought[1] but I'm not currently aware of any better name, except maybe “MarpaX::SLIF::Preprocessor”. I am thoroughly opposed to “MarpaX::CoolName”, but otherwise open to suggestions.
19:57 lwa [1]: https://github.com/latk/p5-MarpaX-Grammar-P​reprocessor/issues/3#issuecomment-134714461
20:01 idiosyncrat_ Re MarpaX::Grammar::Preprocessor -- I started a discussion of its name as a Github issue: https://github.com/latk/p5-Marpa​X-Grammar-Preprocessor/issues/3
20:02 idiosyncrat_ Oops! scrolling error again -- I missed lwa's last.
20:02 idiosyncrat_ Any comments from folks on the channel.
20:03 idiosyncrat_ Re generic names, I know that Parse::Earley was taken when I started Marpa, and that has not proved to be much of an issue.
20:04 idiosyncrat_ Parse::Earley is a very simple implementation of the algorithm exactly as it is in Earley's paper, including the bug.
20:06 idiosyncrat_ Any comments from folks on the channel?
20:06 idiosyncrat_ AFK
22:40 ronsavage joined #marpa
22:52 ronsavage For a short while I thought MarpaX::SLIF::Preprocessor was better than MarpaX::Grammar::Preprocessor, but I now think the latter is better, since we already have a couple of modules under MarpaX::Grammar::*. As for the combined form MarpaX::Grammar::SLIFPreprocessor, it too is tempting but I'd say if other authors want to preprocess grammars, let them use more specific names than MarpaX::Grammar::Preprocessor.
23:27 idiosyncrat_ joined #marpa
23:34 idiosyncrat_ Re MarpaX::Grammar::Preprocessor -- so I guess we have a consensus that MarpaX::Grammar::Preprocessor was a perfectly fine name to begin with :-)

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