Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-01-20

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

All times shown according to UTC.

Time Nick Message
00:39 Idiosyncrat joined #marpa
00:42 Idiosyncrat joined #marpa
00:43 Idiosyncrat VsyachePuz: re http://irclog.perlgeek.de/m​arpa/2016-01-19#i_11910973 -- thanks, it was my most popular blog post.
00:43 Idiosyncrat The book is a mathematical exposition -- it won't be easy reading, but I think that has to be done as a 1st layer.
00:44 Idiosyncrat A history chapter or chapters could be added, but the thing is taking long enough as it is.
02:40 kaare_ joined #marpa
02:48 ilbot3 joined #marpa
02:48 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
03:51 hobbs joined #marpa
05:22 ronsavage joined #marpa
05:42 VsyachePuz Idiosyncrat: i am just saying that list of literature is not sorted by year of publication
05:43 Idiosyncrat What's the AMS standard?
05:45 VsyachePuz Idiosyncrat: I don't know what AMS standart saying about sorting. Does it require to sort authors by alphabet?
05:46 Idiosyncrat I don't either -- but if I'm changing the order of the bibliography, it will be to whatever the AMS standard is.
05:49 Idiosyncrat I think it is date within author, but I cannot find the reference, all I can Google is this -- http://www.math.uiuc.edu/~hil​debr/tex/bibliographies.html
05:52 VsyachePuz thats's flaw of old distribution media - unability to satisfy different use cases. Site can reorder lists at request
06:56 Idiosyncrat Good night!
07:14 VsyachePuz "Mortar" gives an association with weapons. "Calcimine" doesn't
07:16 VsyachePuz or "grout"
07:19 VsyachePuz joined #marpa
11:53 kaare_ joined #marpa
13:51 KotH_ joined #marpa
13:51 btyler_ joined #marpa
13:52 Cheery_ joined #marpa
13:59 kaare_ joined #marpa
14:01 VsyachePuz (*grammar := { [ws], rule }, [ws] ;*)
14:01 VsyachePuz grammar_part1 := (*brick*)ws | empty_sequence ;
14:01 VsyachePuz grammar_part2 := grammar_part1, (*brick*)rule;
14:01 VsyachePuz grammar_part3 := grammar_part2, grammar_part3 | empty_sequence ;
14:01 VsyachePuz grammar := grammar_part3, grammar_part1;
14:02 VsyachePuz here one brick corresponds to 2 places in original rule before rewrite
14:03 JPGainsborough joined #marpa
14:09 Cheery joined #marpa
14:10 VsyachePuz to achieve property "Every brick symbol must translate to exactly one external symbol"
14:11 VsyachePuz i need to make 2 internal rules for each separate [ws] instead of one common rule
14:11 VsyachePuz grammar_part1a := (*brick*)ws | empty_sequence ;
14:11 VsyachePuz grammar_part1b := (*brick*)ws | empty_sequence ;
14:40 VsyachePuz http://irclog.perlgeek.de/m​arpa/2016-01-19#i_11910011 : EBNF standart DOES have ebnf description of itself (in chapter 8 part 8.1 is formal description, part 8.2 is informal description)
14:49 koo7 joined #marpa
16:15 VsyachePuz http://staff.icar.cnr.it/ruffolo/progetti​/projects/10.Parsing%20Earley/2002-Practi​cal%20Earley%20Parsing-10.1.1.12.4254.pdf
17:37 Idiosyncrat joined #marpa
17:38 Idiosyncrat VsyachePuz: re http://irclog.perlgeek.de/m​arpa/2016-01-20#i_11914483 -- ok thanks for catching that.
17:39 Idiosyncrat re http://irclog.perlgeek.de/m​arpa/2016-01-20#i_11912408:
17:40 Idiosyncrat "brick and mortar" is a standard pairing in English, like "salt & pepper".  And IIRC correctly it is used elsewhere in the same sort of figurative sense that I employ it.
17:44 VsyachePuz so, do i understand properly, that I can't avoid duplication of rules in internal grammar?
17:45 Idiosyncrat re http://irclog.perlgeek.de/m​arpa/2016-01-20#i_11914157, the grammar example
17:45 Idiosyncrat You have to attack the problem recursively, one part at a time, so that for the
17:46 Idiosyncrat a ::= b [c] rules
17:46 Idiosyncrat you rewrite to
17:46 Idiosyncrat a ::= b
17:46 Idiosyncrat a ::= b c
17:47 Idiosyncrat Then you have only the sequence to rewrite, and it is in the form
17:47 Idiosyncrat a ::= b*
17:47 Idiosyncrat I solved most of these problems in Kollos, if you want to look at that code -- it's in Lua.
17:48 Idiosyncrat The trick, which I did not completely work out, is what sequence to tackle things in breaking down that
17:48 Idiosyncrat s/that/the/
17:48 Idiosyncrat larger components.
17:48 Idiosyncrat But you've sort of got the idea.
17:50 Idiosyncrat But there are cases where you need to duplicate rules and symbols.
17:50 VsyachePuz My current idea changed. Now I want to rewrite Expression Tree which i got with Eto.Parse parser into Irony parser syntax. That is because I was able to parse EBNF with Eto.Parse's recursive descent parser. And I need my grammer to be written for Irony, because DeveelDB uses Irony.
17:52 VsyachePuz Thus I will be able to write grammar in EBNF and then rewrite it for the LALR parser
17:52 VsyachePuz I don't know is it possible to automate this
17:53 VsyachePuz If this will not work, I will return to implementation of Earley replacement
18:28 ernimril joined #marpa
18:29 ceridwen joined #marpa
18:59 VsyachePuz why do you need to move libmarpa into Koloss? you can try to use git submodule
19:02 Idiosyncrat I spent some time reading up on the git mechanisms for putting one module inside another, and none of them assured me that I would not be painting myself into a corner.
19:03 Idiosyncrat A simple, crude, copy is easy to keep track of.
19:03 Idiosyncrat Eventually the idea is that Kollos will be the next libmarpa, so effectively it will be a copy anyway.
19:06 Idiosyncrat Note, for example, this: https://codingkilledthecat.wordpress.com/2012/04/​28/why-your-company-shouldnt-use-git-submodules/
21:51 ronsavage joined #marpa
22:15 koo7 joined #marpa

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