Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-03-30

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

All times shown according to UTC.

Time Nick Message
01:30 ronsavage1 joined #marpa
01:51 jeffreykegler joined #marpa
06:38 flaviu joined #marpa
06:54 sirdancealot joined #marpa
07:04 ronsavage joined #marpa
08:19 pczarn joined #marpa
09:52 lwa joined #marpa
15:32 sirdancealot joined #marpa
15:38 koo5 joined #marpa
18:37 jeffreykegler joined #marpa
18:45 pczarn joined #marpa
19:13 koo5 joined #marpa
19:41 * jeffreykegler is still working on the strand parsing document
19:42 jeffreykegler The investigation into strand parsing is relevant to Kollos and, despite the time it has taken, I am now glad I decided to look into it first.
19:42 jeffreykegler Because it indicates to me the direction in which I want to take Kollos, and evolve Libmarpa.
19:43 jeffreykegler One feature I plan to add to Libmarpa, is the ability to add rules and symbols to a grammar, even when it is in mid-parse.
19:44 jeffreykegler This won't happen next week or next month, but I want to head in that direction.
19:45 jeffreykegler Note that because of the SLIF's abilities to modify its lexing, and the power of Marpa grammars and LATM, some people get the impression that Marpa *already* can modify its grammar on the fly,
19:46 jeffreykegler but, technically speaking, as of the moment, Libmarpa requires its grammar to be fixed in advance.
20:11 sirdancealot joined #marpa
20:17 jeffreykegler Other stuff that will change -- some Libmarpa features which went nowhere, I expect will be eliminated.
20:18 jeffreykegler Currently, for example, Libmarpa allows not just ambiguous tokens (which have proved *very* useful), but variable length tokens.
20:18 sirdancealot which have proved incomprehensible
20:19 jeffreykegler Variable length tokens were intended to allow unforeseen new lexing techniques.  These techniques never emerged, and making the way for others required me to restrict the upper layers to fixed length tokens.
20:20 * jeffreykegler notes that if incomprehensiblity was the criteria, particularly in the initial reaction, Marpa would simply not exist
20:21 jeffreykegler Another feature likely to disappear is LHS terminals -- traditionally, a symbol is either a terminal or a LHS, never both.  Libmarpa currently does not impose this restriction.
20:22 jeffreykegler But the potential for LHS terminals never did anything but confuse people, and the SLIF, at a higher level, prohibits them.
20:22 jeffreykegler Future versions of Libmarpa may well prohibit LHS terminals at the lower level as well.
20:24 jeffreykegler Anyway, just to let folks know, despite the lack of visible activity, things are taking shape.
20:25 jeffreykegler Thanks!
21:07 ronsavage joined #marpa
21:12 ronsavage jk: Glad yo hear you're so positive.
21:12 ronsavage 'yo' => 'to'.

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