Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-11-30

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

All times shown according to UTC.

Time Nick Message
00:05 ronsavage koo5: Discussion has focused on the fact that asm.js makes things possible. But - does it make things better in some ways? If so, in what ways?
00:06 hobbs not having to ship data, I suppose
00:13 koo5 no idea how to answer that out of the context of a particular project
00:14 koo5 as for me, it means i will be able to make a browser version of my python program easily
00:20 koo5 as for marpa, it means marpa in js
00:21 koo5 im curious how kollos will fare
00:28 koo5 wouldnt something like a referene pseudocode of the high level layer be useful?
02:17 jeffreykegler joined #marpa
02:18 jeffreykegler ronsavage: I talk about it primarily because there is interest in the Marpa community about JS, and I think it best to let people follow their hunches/instincts/enthusiasms
02:19 jeffreykegler My own efforts are focused on the Lua approach, and I encourage folks to join in there.
02:20 jeffreykegler koo5: actually I am writing pseudo-code of the upper layer for Kollos currently
02:21 jeffreykegler I am targeting the Lua effort, but I'm trying to write it to be more generally useful.
02:21 koo5 cool
02:43 ronsavage Yes, I get the interest in JS. I have absolutely no doubt a port of Marpa to JS (however it is done) is inevitable.
04:08 flaviu joined #marpa
09:04 jdurand joined #marpa
09:36 jdurand Re http://irclog.perlgeek.de/​marpa/2014-11-29#i_9734731 - as stated later it was done just because everything that contains the word 'javascript' is eventually turned into gold
09:38 jdurand nevertheless, as stated in the Gary Bernhardt's video JavaScript is a remarquably minor language - and until nowadays, years and years after its creation, most of its community is either doing more or less R&D with it, proving it is not a mature language, or using blackboxes like node, jquery, that are abusing design flows of the language to show a more coherent interface outside of it
09:40 jdurand The fact that JS community is trying to rewrite the whole world once again with it is tiring, and if I had to to a choice, now it would be java 8 - because it now have the only thing IMHO that JS had a metir with: functions closures
09:40 jdurand "had a merit with"
09:43 jdurand Anyway, JS is everywhere of course because it has been marketed, and this cannot be avoided. Playing with asm.js to have a sort-of-and-lying native speed, why not. I'd prefer to let this job to people "playing" with the language - in favour of direct C-to-traget language transpiling - which is a certinaly doable thing when the source is pure C with no, or almost no, dependency. This is what libmarpa is isn't it.
09:49 ronsavage joined #marpa
10:28 koo6 joined #marpa
10:28 koo5 joined #marpa
13:16 flaviu joined #marpa
13:56 lwa joined #marpa
15:54 flaviu joined #marpa
16:11 flaviu joined #marpa
17:34 jeffreykegler joined #marpa
17:34 jeffreykegler https://plus.google.com/10524375​9955561553019/posts/cJs782LvbvM
17:35 jeffreykegler Jean-Damien duplicated a C utility named cdecl using his Marpa-powered C parser.
17:36 jeffreykegler cdecl takes C declarations and tells you what they mean in English.
19:56 koo6 joined #marpa
19:57 koo5 joined #marpa
20:42 idiosyncrat joined #marpa
20:44 idiosyncrat I wrote up a document on how the LUIF will implement evaluation: https://github.com/jeffreykegler/ko​llos/blob/master/design/evaluate.md
20:45 idiosyncrat [ It's about LUIF internals -- the LUIF rewrites the grammar before Libmarpa gets it, and the it must be "un-rewritten".
20:45 idiosyncrat The doc describes how that is done. ]
20:46 idiosyncrat It goes with the document describing the rewrites from a general point of view: https://github.com/jeffreykegler/ko​llos/blob/master/design/rewrite.md
20:47 idiosyncrat At I see a roadmap for Kollos/LUIF leading to a near-term milestone: a JSON parsing implementing in Kollos.
20:48 idiosyncrat I still need to pseudo-code the logic for sequence rule rewrites, but at that point I think we have all the major pieces.
20:48 idiosyncrat With a JSON parser, we can run benchmarks --
20:49 idiosyncrat we will lose something of them since JSON is more lexed than parsed and there are lexers out there hand-optimized in C for it ...
20:49 idiosyncrat JSON::XS contains one.
20:50 idiosyncrat But humility is good for the soul :-)
20:50 idiosyncrat and it will tell us where we stand.
20:51 idiosyncrat [ Kollos does not yet have a lexer.   For the Kollos first cut at a JSON parser, we can write (steal?) a lexer written in Lua. ]
21:17 btyler joined #marpa
22:08 koo6 good stuff, i will have a ton of questions when i get back to marpa, but with those docs i will at least have an idea what to ask
22:09 flaviu1 joined #marpa
22:49 jeffreykegler joined #marpa
23:05 ronsavage joined #marpa

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