[02:53:01] vin-ivar: "the $ character representing the empty string"? Is their regex notation idiosyncratic? I can't get any google hits for "the $ character representing the empty string" except that site. [03:01:56] And I can't see how, for example, 'b+a*aa' will match 'b', where '+' is the choice operator. [03:02:43] What does choice operator mean? the same as unix regex '?'? [04:12:36] *** Joins: Gallaecio_ (~quassel@84.120.183.107.dyn.user.ono.com) [04:12:58] *** Parts: Gallaecio_ (~quassel@84.120.183.107.dyn.user.ono.com) () [04:26:21] *** Quits: spectei (~fran@unaffiliated/spectie) (Ping timeout: 244 seconds) [08:56:26] *** hellocomputer is now known as eerbmt [08:58:26] drbean: I think it's like | [08:58:34] so (b|a)*aa [08:58:48] but yeah it shouldn't match b in that case [09:03:58] *** Joins: juanfran (c191e609@gateway/web/freenode/ip.193.145.230.9) [09:05:05] *** Quits: juanfran (c191e609@gateway/web/freenode/ip.193.145.230.9) (Client Quit) [09:05:36] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [09:19:29] *** Joins: spectei (~fran@ilazki.thinkgeek.co.uk) [09:33:37] *** Joins: memduh (uid173783@gateway/web/irccloud.com/x-vgsfkhdjhidqihxo) [10:02:18] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [10:04:39] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [10:12:43] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [10:43:33] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [10:51:30] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [11:33:27] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [11:52:27] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [11:55:48] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [11:58:01] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Client Quit) [12:02:33] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [12:05:31] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Client Quit) [12:12:26] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [12:15:54] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Client Quit) [12:34:44] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [12:48:27] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [12:48:33] inariksit, is there a paper talking about bind and soft bind and stuff [12:48:42] i think krasimir had one? [12:53:58] I don't remember, but it's pretty recent, so try anything post-2014 [12:55:14] memduh: about the John loves Mary, the marking of the object doesn't have to be in the abstract syntax [12:55:37] aarne has these examples where Latin has a case, and then NP has a lincat {s : Case => Str} [12:55:43] *** Joins: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) [12:55:45] and then you choose the right case in the PredVP function [12:56:00] now, if it's just a preposition, the type of the NP can still be just {s : Str} [12:56:12] but the PredVP function adds just a hardcoded preposition [12:56:28] should there be some sort of function NP->NP that adds the preposition? [12:56:42] I thought of that but it seemed hackish but im not sure [12:56:57] not necessary, it's fine to have it added by the Pred function [12:57:13] ah so the v2 should add it a a string? [12:57:14] of course in the actual resource grammar, if it depends on the verb which preposition to use, you have to add a parameter into the verb [12:57:21] I see [12:57:45] in the actual grammar then, it cant be the same function for the two things? Because they have different arguments [12:58:06] what do you mean by "the two things"? [12:58:18] the complement in english is loves mary [12:58:25] Compl np v2 [12:58:34] in kurdish it will be pr np v2 [12:58:35] ah sorry yeah I meant Compl, not Pred [12:58:45] so we can overload the compl like that? [12:59:18] for John loves Mary, you only have one verb, so you can just have the Compl function return "specific pr" ++ np ++ v2 [12:59:43] in a more general grammar, your Compl can return ++ np ++ v2 [13:00:11] or the field in v2 can simply be a string :) so it could be v2.pr ++ np ++ v2 [13:00:12] ah thats good, I thought the return had to be the same as the abstract [13:00:23] hmm [13:00:28] that makes sense honestly [13:00:32] ^^ [13:01:16] in GF rules, you can freely add more stuff to the right hand side; you can duplicate some argument on the left hand side; or you can omit some argument on the left hand side [13:01:42] so you can perfectly fine write linearisations like Compl v2 np = "I am a VP!" [13:01:55] which is of course a crappy grammar :-P [13:02:12] :D [13:02:44] then whenever you parse "I am a VP!", you get bunch of question marks in your tree [13:02:56] cause it doesn't know which V2 or which np is it using [13:03:04] D: [13:03:08] *** Quits: juanfran (c191e609@gateway/web/cgi-irc/kiwiirc.com/ip.193.145.230.9) (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) [15:10:02] memduh: https://github.com/GrammaticalFramework/gf-contrib/tree/master/miniresource [15:10:25] you can get the abstract syntax from there, and just copy some existing into MiniresourceKur, or start from scratch [16:55:55] *** Quits: spectei (~fran@ilazki.thinkgeek.co.uk) (Ping timeout: 258 seconds) [22:56:02] *** Joins: spectei (~fran@unaffiliated/spectie) [23:53:12] *** Quits: spectei (~fran@unaffiliated/spectie) (Ping timeout: 276 seconds)