Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-10-14

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

All times shown according to UTC.

Time Nick Message
01:32 Idiosyncrat ronsavage: re http://irclog.perlgeek.de/m​arpa/2015-10-13#i_11370516 -- was it actually an L0 lexeme?
01:33 Idiosyncrat Marpa has a specific definition of L0 lexeme, meaning a symbol which appears as the LHS of an L0 rule, but *not* on the RHS of any L0 rule.
02:50 ronsavage Ah! I does appear on the RHS of an L0 rule. FAQ here we come.
02:51 ronsavage In other news, I've just re-written Genealogy::Gedcom::Date to use Marpa. The docs will be updated over the next few days.
02:59 Idiosyncrat Something perhaps to note in the FAQ -- "lexeme" is a technical term within Marpa ...
02:59 Idiosyncrat it's not intended to mean what you intuitively might think of as a lexeme.
03:00 Idiosyncrat It's for that reason I chose the term "lexeme", which is unusual, as opposed to "token" -- I was trying to convey the idea that this is a special technical Marpa-specific term ...
03:01 Idiosyncrat which is defined in the docs.
03:02 Idiosyncrat It seems that the term "lexeme" has become more widespread so that it now has the kind of intuitive overload I was trying to avoid.
03:03 Idiosyncrat I half-wonder if Marpa popularized it, but I have to think it more likely I underestimated how widely used the term actual is.
03:25 ceridwen https://www.google.com/trends/explore#q=lexeme
03:26 ceridwen FWIW, I learned the term in linguistics, and I think it's fairly old.
03:27 ceridwen http://www.etymonline.com/index.php?term=lexeme
03:27 Idiosyncrat Linguistics is where I got the term from -- I expected that it would be rare outside the technical linguistics context
03:30 ceridwen https://en.wikipedia.org/wi​ki/Lexical_analysis#Lexeme
03:31 ceridwen Apparently it's also mentioned in the 2nd ed. of the canonical compilers text: https://stackoverflow.com/questions/149​54721/what-is-the-difference-between-to​ken-and-lexeme#comment20999371_14958865
03:33 Idiosyncrat That may be my blind spot -- the best best in the "dragon series" IMHO was the first edition.
03:33 Idiosyncrat I had given up on the series by the time the purple book came out.
03:34 Idiosyncrat s/best best/best book/
05:21 ronsavage joined #marpa
05:37 ronsavage joined #marpa
08:31 ronsavage joined #marpa
08:36 sivoais joined #marpa
08:39 ronsavage I've just been searching the backlog. Wasn't there a comment that the terminology 'L0' should be abandoned? I can't find any reference to that.
08:42 ronsavage I've updated the FAQ re that gist/error msg. See q32 at http://savage.net.au/Perl-mod​ules/html/marpa.faq/faq.html.
08:46 sivoais joined #marpa
08:57 sivoais joined #marpa
09:07 sivoais joined #marpa
09:17 sivoais joined #marpa
09:27 sivoais joined #marpa
09:37 sivoais joined #marpa
09:48 sivoais joined #marpa
09:48 sivoais joined #marpa
09:58 sivoais joined #marpa
10:08 sivoais joined #marpa
10:19 sivoais joined #marpa
10:29 sivoais joined #marpa
10:39 sivoais joined #marpa
10:49 sivoais joined #marpa
10:59 sivoais joined #marpa
14:22 Idiosyncrat ronsavage: re the 'L0' name -- no, I'm not thinking of abandoning it, at least not in Marpa::R2
14:23 Idiosyncrat It is, yes, based on a scheme for multiple grammars/lexers which I abandoned, but ...
14:23 MerittClarksdale joined #marpa
14:24 Idiosyncrat it has proved to be one of my more successul naming conventions -- people seem to instantly grasp what it means.
14:25 Idiosyncrat The success comes from luck rather than skill but, hey, I'll take it.
14:25 Idiosyncrat ronsavage: re the new FAQ entry -- thanks!
14:32 MerittClarksdale Idiosyncrat, are you fond of PL/I
14:33 Idiosyncrat I haven't used it in many, many years
14:33 Idiosyncrat Not since 1976 or so
14:33 Idiosyncrat It was historically important -- introduced many innovations
14:34 MerittClarksdale Yes, we have some legacy apps here written in PL/I with CICS on MVS
14:34 Idiosyncrat In some ways kind of the Perl of its time -- though without the same success
14:35 Idiosyncrat But I would not choose it for new code without some organizational/economic reason that really indicated it was the way to go.
14:36 Idiosyncrat I don't think there's an open source compiler for it.
14:36 Idiosyncrat Legacy apps?  How old?
14:37 MerittClarksdale right, the GNU people tried but gave up, the hard part perhaps wasnt the language itself but their idea of trying to work with the GCC backend
14:37 MerittClarksdale They probably would have had better luck generating C code as GNUCobol does
14:38 MerittClarksdale Our apps date back 30-40 years
16:40 btyler_ joined #marpa
16:59 djns joined #marpa
16:59 Cheery joined #marpa
17:07 lucs joined #marpa
17:07 spacebat joined #marpa
17:08 shadowpaste joined #marpa
17:10 djns joined #marpa
17:10 Cheery joined #marpa
17:29 Idiosyncrat joined #marpa
21:23 ronsavage joined #marpa
21:23 ronsavage MerittClarksdale: I always felt PL/I was a type of lock-in. IBM decided to offer a language that have everything in it which you'd ever need, and hence you'd never have a reason to use another language (as well). Baroque in complexity, you might say.

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