Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-11-03

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

All times shown according to UTC.

Time Nick Message
01:10 Idiosyncrat joined #marpa
01:12 Idiosyncrat joined #marpa
01:13 Idiosyncrat ronsavage: re http://irclog.perlgeek.de/marpa/2015-11-02#i_11471264
01:13 Idiosyncrat One good aspect of Marpa's initial reception.
01:14 Idiosyncrat I've been able to work out this line of inquiry by myself, without any pressure, and without anyone or anything breathing down my neck.
01:15 Idiosyncrat It's been an unbelievable privilege.
01:55 Idiosyncrat joined #marpa
04:58 ronsavage joined #marpa
05:20 lucs joined #marpa
05:39 lucs joined #marpa
07:52 ronsavage joined #marpa
09:29 ronsavage JK: Sounds good, but lack of support can be disheartening. Nevertheless, you've clearly managed....
13:22 Pursuit joined #marpa
15:40 user2_ Idiosyncrat, what is your view regarding the Perl 6 debacle, its musical VMs and implementations (Parrot, JVM, Moar, etc).
15:41 Idiosyncrat I would resist calling it a debacle
15:42 Idiosyncrat Of course, I, like everyone else, had hoped to see Perl 6 come out long ago.
15:43 Idiosyncrat But at worst, perhaps Perl 6 will be the ALGOL of our generation -- ALGOL was never much used --
15:43 Idiosyncrat except by ronsavage :-)
15:43 user2_ Some have said Parrot should have focused on supporting Perl 6, rather than trying to be the one VM to rule them all
15:43 Idiosyncrat but it was hugely influential.
15:44 Idiosyncrat I did wonder at the time about the direction of the Parrot team.
15:44 Idiosyncrat I considered, early on, btw, basing Marpa on Parrot, and contacted them.
15:45 Idiosyncrat I got some very intelligent feedback from the team, but ...
15:46 Idiosyncrat they were very focused on OO innovation -- to the point where if you wanted to do structures using Parrot, ...
15:46 Idiosyncrat that is pointer plus offset
15:46 Idiosyncrat they considered you hopelessly archaic
15:47 Idiosyncrat On the other hand, if they had revolutionized OO, we'd all be aclaiming the genius of that move now ...
15:47 Idiosyncrat and it is very hard to call these things.
15:47 Idiosyncrat But I think many projects have been the victims of a fetish for OO experiments and Parrot was one.
15:49 user2_ I tried to look into Parrot once, and found the documentation to be unclear. The good things about perl 5 that make it inviting is the extensive documentation of the language and the centralized package repository. Perl 6 suffered from no official implementation which for people investigating it lead to confusion and perceptions that it was not ready.
15:49 Idiosyncrat And I really do think that their early distancing from Perl 6 was a mistake
15:49 user2_ right
15:49 Idiosyncrat In Marpa, I made it a rule, find users and *support* them as the first priority.
15:50 Idiosyncrat Because users teach you things about what works and what does not ...
15:50 Idiosyncrat and what will be useful and what will not.
15:51 Idiosyncrat In the early days, I got a lot of contacts re Marpa wanting to take it in the direction of OO experimentation ...
15:51 Idiosyncrat which I resisted
15:52 Idiosyncrat ... in a friendly way, that is.  If somebody wanted to go ahead with their Marpa-powered OO project, I'd support them.
15:53 Idiosyncrat But I let my own focus be elsewhere -- and in fact, nothing ever came of those suggestions.
15:53 Idiosyncrat It's not that the combination of OO and Marpa might not be a great idea -- it's just that the record of wasted time is *so* bad.
15:53 Idiosyncrat And Parrot is Exhibit A.
15:54 Idiosyncrat But guessing how to take this kind of project is hard.
15:54 Idiosyncrat Marpa is a success, not because I was 100% accurate, ...
15:55 Idiosyncrat but because I keep at it, and my persistance compensated for the dead ends I pursued.
15:55 Idiosyncrat One thing I will say about Perl 6 ...
15:56 Idiosyncrat and I want to emphasize I am *not* joining in with the critics of Perl 6, I am talking only about one point ...
15:57 Idiosyncrat is that I do not think Perl6 grammars have the horse-power to deliver on their vision of Language-oriented-programming.
15:58 Idiosyncrat But Marpa has benefited from its association with the Perl 6 community.
15:59 Idiosyncrat When I did SLIF features, the first thing I would do is study what Larry was doing in Perl 6.
15:59 Idiosyncrat So Marpa is heavily Perl 6 influenced, though it would be hard to tell that from its outward look.
16:01 Idiosyncrat And whatever happens with Perl 6, we need to ask ourselves this question"
16:01 user2_ There were said to be many very innovative features with perl 6, my design philosophy is Mechanism Not Policy-Recommended Practices. I do not put artificial limits on how my software can be used, I make recommendations on how they could be used. But I cannot foresee someone elses needs. Perl 6 might be considered a toolbox of features, that some may need and others not
16:01 Idiosyncrat "How many times in a lifetime does Larry Wall need to prove that he is a genius?"
16:04 Idiosyncrat user2_: yes, and Larry Wall's toolkit philosophy was immensely influential on me.
16:04 Idiosyncrat I tried to put features in Marpa that could be used in ways that I did not forsee.
16:05 user2_ Yes. Thats what we like about Perl 5, one of many things. Its a toolbox of features that tries to avoid creating a straightjacket environment and letting the programmer decide how to do it, and giving the programmer many ways to do it
16:06 Idiosyncrat Yes, and Perl 5 was extravagantly ambitious, by the standards of the time, but Larry pulled it off.
16:06 user2_ perls TIE interface had an influence on me, showing how when adding a facility, by making it one more generalized one can allow the facility to be used in other ways other than initially designed for.
16:06 Idiosyncrat TIE?
16:07 user2_ tie'd variables, sorry
16:07 user2_ tie()
16:09 Idiosyncrat By the way, the toolkit philosophy comes out of UNIX and its best writeup ...
16:09 Idiosyncrat is Kernighan & Pike _The Unix Programming Environment_
16:10 Idiosyncrat This book, unfortunately, is so old that it used a per-C89 dialect of C, which means it is hard to recommend it to a younger programmer.
16:12 Idiosyncrat When I read it, it was all code that I could use directly.
16:13 Idiosyncrat But I am told there are updated examples for it out on the web somewhere.
16:17 user2_ right
16:25 Idiosyncrat Btw, re guessing which way to go with a projects, one thing I did get right was identify what I call the 3 Temptations of Marpa
16:25 Idiosyncrat http://irclog.perlgeek.de/marpa/2015-05-19#i_10624305
16:25 Idiosyncrat 1.) Worrying about whether an optimization is premature.
16:26 Idiosyncrat 2.) Multi-threading
16:26 Idiosyncrat 3.) OO
16:26 Idiosyncrat All of these are actually basically good things, but they have been vastly overemphasized,
16:27 Idiosyncrat especially when it comes to leading-edge projects.
16:27 Idiosyncrat so overemphasized that they wreck projects.
16:39 user2_ good points yes
16:45 Idiosyncrat AFK
19:47 Idiosyncrat joined #marpa
21:08 ronsavage joined #marpa

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