Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-12-10

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

All times shown according to UTC.

Time Nick Message
00:39 idiosyncrat_ joined #marpa
01:08 koo7 joined #marpa
02:30 ronsavage joined #marpa
02:34 ronsavage JK: Perhaps it's still early enough in Perl 6's evolution that you could influence their parser's design. For that reason, I'd suggest saying something. It's them up to them as to what they do and how they treat you.
02:34 ronsavage And how they treat you affects how we respond. I could work in casting Perl 6's grammar into SLIF (I have some time), but not if they're nasty toward you.
02:47 ilbot3 joined #marpa
02:47 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
03:27 Idiosyncrat joined #marpa
03:27 Idiosyncrat When it comes to nastiness, the Perl 6 community has been mostly the recipient, and their reaction has been to turn the other cheek.
03:28 Idiosyncrat Also, they have been great patrons of Marpa, and helpful with advice.
03:28 Idiosyncrat So if a certain amount of snark came my way from them, I'd think I owed it to them to take it gracefully.
03:29 Idiosyncrat However, I do think that advent post sends folks on a "wild goose chase", and I should make some kind of response.
03:30 Idiosyncrat I'll note I felt a similar kind of dilemma with PEG -- I didn't want to criticize it directly ...
03:31 Idiosyncrat PEG's problems became painfully apparent if you try to use it, and I thought I could just let things take its course.
03:31 Idiosyncrat But I consistently underestimate people's tolerance for pain. :-)
03:32 Idiosyncrat I'm planning to start with a postive, supportive post of Perl 6, just to set the tone correctly.
03:32 Idiosyncrat I have a draft of one, that I will share with the channel over the next 24 hours, to get preliminary reaction before I put it out there.
03:33 Idiosyncrat In particular, I will appreciate feedback on whether others will see it as having the same positive tone that I read into it and intend for it.
03:34 Idiosyncrat ronsavage: but re the Perl 6 BNF.
03:34 Idiosyncrat They've gone 13+ years without bothering to create BNF, something I'd have started on in the first 13 hours.
03:35 Idiosyncrat I think we've a real divergence of mindset here, so much so that creating BNF for them would be pointless --
03:35 Idiosyncrat if they don't see its value, they don't see its value.
03:37 Idiosyncrat So while I think we owe them not just tolerance, but gratitude, there is no much point in doing something for them that they feel they simply do not have a use for.
03:38 Idiosyncrat As for influencing the design of the Perl 6 parser -- this is something I've thought about long and hard for years.
03:39 Idiosyncrat If you study the Perl 6 design docs, you see they've committed heavily to top-down parsing
03:39 Idiosyncrat Of course, Marpa did not exist 13 years ago, though I'm not sure that would have made any difference in Larry's decision.
03:40 Idiosyncrat It's kind of like advising a friend re his choice of wife.
03:40 Idiosyncrat If he's just met her, you can drop a hint.
03:40 Idiosyncrat But after he's committed, if you want to be a friend, you just shut up about your problems.
03:41 Idiosyncrat So I've avoided raising the question of the parsing underlying Perl 6 -- believe me, they are committed at this point.
03:43 Idiosyncrat However, if they are promoting Perl 6 grammars for other things, the way the advent article was, that's different ...
03:43 Idiosyncrat and there I feel justified in pointing out if they can't do the job.
03:44 Idiosyncrat Anyway, I'll paste a draft of my post on Perl 6 over the next 24 hours, ...
03:45 Idiosyncrat which will not talk about parsing, but which will be an apprecaition of their effort from a wider perspective.
03:45 Idiosyncrat And I will very much be interested in the comments of folks on this channel.
03:54 ceridwen I read through the Perl 6 grammar post briefly (my useful thinking time is kind of limited these days, it's annoying).  My take on it is that it looks a lot like parser combinators implemented with classes.  I like parser combinators, I'm currently planning to use them as the API in my project, but they're honestly not a big deal to implement in any language I've tried it in.
03:57 ceridwen They don't seem to address the parsing implementation at all in their post, which is the part that might be hard to implement and describes the performance of the algorithm.
03:58 ceridwen Idiosyncrat, Feel free to criticize away at least on stuff I'm working on, I learn new things from well-founded criticism.
03:59 ceridwen On another topic, though, Afroozeh and Izmaylova published this in like Oct: https://cdn.rawgit.com/iguana-parser/papers/master/onward15.pdf
04:00 Idiosyncrat joined #marpa
04:00 ceridwen Which contains a lot of the same ideas I've been working on, GLL with data-dependent grammars.
04:02 ceridwen I agree with your take on PEGs---they look promising, but on closer inspection the packrat parser's linear performance is bought with an awful lot of memory (the O(n^3) algorithms take a lot of memory too, but they run in O(n) on many languages) and prioritized choice introduces all sorts of problems into writing grammars.
04:03 ceridwen I'd still like to investigate the use of PEG-like features for pattern matching some day, but I don't think they're very good for parsing in general.
04:10 ronsavage JK: OK. Thanx for the extended reply. My ignorance of the Perl 6 state-of-the-art is there for all to see, but it does not worry me :-).
12:10 CQ joined #marpa
16:26 idiosyncrat_ joined #marpa
16:30 idiosyncrat_ https://gist.github.com/jeffreykegler/72cfb4b41343c5731821
16:31 idiosyncrat_ The "Appreciation of Perl 6" I promised last night.
16:32 idiosyncrat_ I will be very interested in your thoughts & reactions.
16:36 lucs Who's the intended audience?
16:37 lucs (Because SLIF is mentioned without context)
16:37 lucs ( s/a interface/an interface/ )
16:52 lucs The last two paragraphs sound like you expect Perl 6 usage to be low, did you mean to convey that?
16:54 lucs ( not sure what the "...Ned into prominence." sentence is doing there )
17:02 lucs "Larry and I have committed to very different approaches" I want to read more! :-)  (I know, later blog post)
17:04 idiosyncrat_ lucs: thanks for the feedback, which is of exactly the kind I wanted.
17:04 idiosyncrat_ I don't have a prediction about the future usage of Perl 6.
17:05 idiosyncrat_ It'd certainly be nice to see a lot of it.
17:06 idiosyncrat_ But I certainly do *not* wish to convey I expect the usage to be low.
17:07 idiosyncrat_ The intended audience is those who care what I think about Perl 6. :-)
17:08 idiosyncrat_ And the intent is to set a positive tone, because most of what I have to say about Perl 6 grammars as a road to meta-programming, 2nd order languages, etc., will be very negative.
17:08 lucs This is for your blog, right? So I guess readers there will know (or find out easily) what the SLIF is.
17:09 idiosyncrat_ The intent here is not necessarily to attract a lot of attention, but just to get my overall perspective on Perl 6 on record.
17:10 idiosyncrat_ I haven
17:11 idiosyncrat_ 't before, because they was no particular reason to think folks cared what I thought of Perl 6. :-)
17:11 idiosyncrat_ But people do ask now, and I think more will ask (or make false conjectures) in the light of any blog posts I write on Perl 6 grammars.
17:13 idiosyncrat_ I want to be positive, but not gushy or insincere -- nobody knows what the take-up of Perl 6 will be.
17:14 idiosyncrat_ My idea in general is that the more attention Larry Wall's work gets, the better a world this becomes.
17:15 idiosyncrat_ But that is not what drives the TIOBE numbers.
17:16 idiosyncrat_ As for where I put it, maybe my blog, maybe their advent calendar, if it's not full by the time I ask.
17:16 idiosyncrat_ I have been asked regularly to contribute to their advent calendar, though I was not this year, I assume because I've always said "too busy" before.
17:16 lucs I'm not as fluent as I'd like in English, so I'll take a couple of minutes to write up what I want to say and come back to paste it here :)
17:18 idiosyncrat_ By the way, in terms of avoiding predictions about take-up, and not expecting technical merit will drive large numbers ...
17:18 idiosyncrat_ this is nothing that I would not also say about Marpa.
17:19 idiosyncrat_ The only difference is that Larry and his Perl 6 team are sure to go into the history books on languages ...
17:19 idiosyncrat_ and while I'd like an equal or better place for those folks who have helped me with Marpa ...
17:19 idiosyncrat_ it is too early to say.
17:21 idiosyncrat_ You all surely deserve it in my opinion.
17:24 lucs I've been learning Perl 6 in the last week or so, and it's incredibly satisfying.
17:24 lucs They've done amazing work.
17:27 idiosyncrat_ And as to the length of time Perl 6 has taken, I'd point out ...
17:28 idiosyncrat_ that when I want project management advice on Marpa, I go over to #perl6, and ask if Larry is around.
17:28 idiosyncrat_ You don't hear Larry asking me for advice. :-)
17:35 lucs :)
17:43 lucs So, those last two paragraphs sound to me like: - We can't know how much usage Perl 6 will get in the future. - High quality (hQ) does not imply high usage (hU). - Earlier text implies that Perl 6 appears to be of high quality (P6hQ), but it's a too easy and too frequent fallacy to infer that (hQ ¬⇒ hU ∧ P6hQ) ⇒ ¬P6hU. - You warn about the mistake of looking only at (sounds like an implied low) level of usage.
17:43 lucs I believe this makes it feel like you expect low future usage (although you don't say that at all).
17:43 lucs Not sure exactly how to make it sound better though :-)
17:47 lucs (I hope that Unicode came out right in your IRC client)
17:47 * lucs & # Lunch!
17:52 idiosyncrat_ I think the "meritocratic fallacy" is very common among programmers --
17:52 idiosyncrat_ The meritocratic fallacy: quality work => success.
17:53 idiosyncrat_ And if you accept that, then you accept the contrapositive --
17:53 idiosyncrat_ that lack of success shows that the work was not quality work.
20:44 ronsavage joined #marpa
20:51 ronsavage And don't forget Burroughs wrote their operating systems in their own hyper-extended version of Algol, which they called Espol. And their own (basic) version of Algol (not Espol) was joking called Over-extended Algol. By that token, their Algol was massively successful. I wrote a tiny bit of Espol, and 13 years of Algol. It was a exceptional and wonderful opportunity.
20:55 ronsavage JK: I see the post as being extremely complimentary about Larry, and about his efforts and contributions. It doesn't actually say much about Perl 6, but it's clear you're not giving it any great praise. And that's fine. You should not feel compelled to.
20:58 pczarn joined #marpa
21:01 idiosyncrat_ OK.  It's pretty clear that piece is not getting across my high estimate of Larry's work and Perl 6's potential.
21:02 idiosyncrat_ I've been studying it for years, but in the past year I've put other things on the front burner.
21:02 idiosyncrat_ (Yes, the Marpa book is coming along.)
21:02 idiosyncrat_ So soon most of you will know more about Perl 6 than I do.
21:02 idiosyncrat_ But I hope to catch up.
21:03 idiosyncrat_ Thanks for allowing me to use you as a "focus group".
21:03 idiosyncrat_ I can't make that piece public as it is -- it will clearly not be read in the way I intended.
21:04 idiosyncrat_ For those who wonder about my carefulness here and my perplexity.
21:05 idiosyncrat_ There's been a lot of nasty Perl 6 bashing, and I don't want to be associated with it -- Perl 6 is not above criticism, but nastiness is counter-productive.
21:06 idiosyncrat_ When I started Marpa I considered it part of the Perl 6 effort, and in a way I still do.
21:08 idiosyncrat_ In fact, there is a Marpa Perl 6 port, so Marpa is now more part of the Perl 6 effort than ever.
21:10 idiosyncrat_ I do feel that I need to respond to this: https://perl6advent.wordpress.com/2015/12/08/day-8-grammars-generating-grammars/
21:11 idiosyncrat_ Which in my frank and humble opinion just invites people down a road that leads nowhere.
21:15 idiosyncrat_ I'
22:39 CQ idiosyncrat: just read the piece. A few comments:
22:40 CQ - first two paragraphs: you're stating respect for larry and hiw work.
22:41 CQ - the next few are an ALGOL excursion, with the main point being that people wanted to use it for something the designers hadn't anticipated.
22:42 CQ - there is a huge jump from the third to last paragraph to the last two paragraphs
22:43 CQ Suggestion: flip the last two paragraphs, and build a bridge from the ALGOL-Ned-prominence-paragraph to the Level-of-usage sentence. That will connect it better and make it flow
22:45 CQ Suggestion2: I would add something to the end of the Usage paragraph, or add a last paragraph with direct thoughts on Perl6: it's a new language, heading into new territory, it had a long start, but has potential because of a,b,c, or something like that.
22:46 CQ currently the sentiment of the final two paragraphs is a tad negative, which I'm sure is not intended
22:47 CQ hope this helps...

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