Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-11-27

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

All times shown according to UTC.

Time Nick Message
02:59 ilbot3 joined #marpa
02:59 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Code paste/run: https://f.perlbot.pl/#marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today - Youtube channel: https://www.youtube.com/channel/UCYKVfGBtfTqbs1JdYq-dc5g
07:54 idiosyncrat joined #marpa
07:54 idiosyncrat Demat!
07:54 idiosyncrat left #marpa
16:06 idiosyncrat joined #marpa
16:09 idiosyncrat https://metacpan.org/pod/release/JKEGL/Marpa-R2-4.000000/pod/Scanless/DSL.pod#null-ranking
16:09 idiosyncrat I am thinking of dropping "null ranking" in Marpa::R3.
16:09 idiosyncrat I assume most of you don't even remember what it is.
16:10 idiosyncrat My question: is anyone using it?
16:10 idiosyncrat My conjecture is that nobody uses it except for me, in the R2 test suite.
16:11 idiosyncrat Null ranking, btw, allows you to designate nulled symbols as "high rank" or "low rank", and rules with identical symbols, but where one is nulled and another is not, will be ordered accordingly.
16:12 idiosyncrat Its effect can be accomplished, perhaps less opaquely, by rewriting the rule into variants where the nulling is explicit, and ranking the variants.
18:04 idiosyncrat null-ranking is obscure, arbitrary, limited, was never requested by a user and AFAIK was never used by an app.  So you might ask, "Why is it in there at all?"
18:04 idiosyncrat The answer is the my very first version of Marpa used the example grammar in the Aycock-Horspool paper which was
18:05 idiosyncrat something like
18:05 idiosyncrat Top ::= A A A A A
18:05 idiosyncrat A :: 'a'
18:06 idiosyncrat Oops make that
18:06 idiosyncrat A ::= 'a'
18:06 idiosyncrat A ::=
18:07 idiosyncrat I wanted to test ordering rules, and since I only had one rule and the only thing was varied was the pattern of nulls, I invented null ranking.
18:21 idiosyncrat left #marpa
22:01 ronsavage joined #marpa
22:03 ronsavage JK:: I only ever used high_rule_only. In MarpaX::Languages::Perl::PackUnpack I have this comment: # BTW: These ranks only work because I've used (ranking_method => 'high_rule_only'). Will you work affect me?
22:04 ronsavage 'you' => 'your'. I ask because such esoterica is beyond me and you did mention 'high' :-).
23:17 idiosyncrat joined #marpa
23:17 idiosyncrat ronsavage: re https://irclog.perlgeek.de/marpa/2017-11-27#i_15501007
23:18 idiosyncrat I plan to keep 'high_rule_only'.  If something uses it breaks let me know, I'll almost certainly treat it as a bug.
23:21 idiosyncrat 'high_rule_only' is, I expect, very widely used.  I like to encourage aggressive use of ambiguity in Marpa and, for ambiguous grammars, there are some important thing you can't reasonably do in any other way.
23:21 idiosyncrat If 'high_rule_only' did go, it would only be to replace it with something at least as powerful and more convenient.

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