[00:17:07] *** Quits: spectie (~fran@unaffiliated/spectie) (Ping timeout: 272 seconds) [00:28:45] *** Joins: spectie (~fran@unaffiliated/spectie) [02:41:58] *** Quits: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) (Ping timeout: 264 seconds) [02:47:46] *** Joins: airini (~airini@253.Red-88-26-239.staticIP.rima-tde.net) [02:50:50] *** Quits: airini (~airini@253.Red-88-26-239.staticIP.rima-tde.net) (Client Quit) [02:59:31] *** Joins: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) [06:43:05] *** Quits: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) (Quit: Leaving) [08:22:33] *** Joins: myoriel (~Myoriel@aftr-88-217-181-187.dynamic.mnet-online.de) [08:56:58] *** Quits: myoriel (~Myoriel@aftr-88-217-181-187.dynamic.mnet-online.de) (Ping timeout: 264 seconds) [09:41:02] Mmmh. 'let' appears to be a keyword. [11:36:15] *** Quits: spectie (~fran@unaffiliated/spectie) (Read error: No route to host) [11:41:29] *** Joins: spectie (~fran@213.87.130.165) [11:41:29] *** Quits: spectie (~fran@213.87.130.165) (Changing host) [11:41:29] *** Joins: spectie (~fran@unaffiliated/spectie) [11:50:38] drbean: yep, you can do local definitions with it [11:50:39] http://www.grammaticalframework.org/doc/gf-reference.html#toc5 [13:10:25] *** Joins: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) [13:15:49] *** Quits: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) (Quit: Leaving) [13:15:58] *** Joins: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) [13:22:59] *** Quits: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) (Quit: Leaving) [13:23:08] *** Joins: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) [14:18:38] *** Joins: jstar (~jstar@unaffiliated/jstar) [14:19:35] *** Quits: spectie (~fran@unaffiliated/spectie) (Ping timeout: 252 seconds) [14:30:19] *** Quits: evariste (~evariste@2001:4b98:dc0:41:216:3eff:fef8:98d6) (Ping timeout: 252 seconds) [14:30:41] *** Quits: MasX (~masx@unaffiliated/masxmasx) (Ping timeout: 252 seconds) [14:31:00] *** Joins: MasX (~masx@2001:41d0:a:2340::1) [14:33:03] *** Joins: evariste (~evariste@2001:4b98:dc0:41:216:3eff:fef8:98d6) [14:40:38] *** Quits: MasX (~masx@2001:41d0:a:2340::1) (Changing host) [14:40:38] *** Joins: MasX (~masx@unaffiliated/masxmasx) [15:31:22] *** Quits: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) (Quit: Leaving) [16:22:51] *** Joins: brinerustle (~Brian@250.Red-88-14-187.dynamicIP.rima-tde.net) [16:27:21] *** Joins: spectie (~fran@unaffiliated/spectie) [19:12:30] hi brinerustle! have you had any progress in your project? [19:13:26] inariksit: hi, thanks for asking! [19:13:43] hey inariksit ! [19:13:45] not with gf, couldn't wrap my head around it at all, i'm afraid. [19:13:55] hey spectie :) [19:13:58] brinerustle, that's what #gf is for! [19:14:03] brinerustle: aww :P but with something else then? [19:14:15] most important that you find something that helps you get stuff done [19:14:17] but it occurred to me brinerustle that you might be as well just doing glosses [19:14:20] but it think i figured out more or less what to do with apertium [19:14:25] e.g. you don't really need "translation" [19:14:36] brinerustle: ok, cook :) [19:14:56] ... [19:14:57] cool [19:14:58] :D [19:15:19] well, it's still in experiemental phase, so not sure what the final output I want is, whether to translate morpheme for morpheme, leave parts untranslated, etc. [19:15:37] although inari's gf output really looked like something that would work too [19:16:13] yes, but I want to be able to work with html, not sure if gf would work, i need an untagger and a tagger [19:16:29] ah, you mean formatting ? [19:16:33] detagger I think it's called, sorry [19:16:41] deformatter / reformatter [19:16:44] yes, that is quite convenient with apertium [19:16:45] yes [19:16:56] morpheme for morpheme sounds like apertium works better for that [19:17:00] gf might be a lot easier but i'm not wrapping my head around it [19:17:08] and yep, everything about convenience [19:17:15] wrt. different input formats [19:17:29] like, with gf you should probably write your own script to normalise the text [19:19:41] and apertium just handles stuff, doesn't matter if you have wEiRd caPITaliSaTioN or space or no space with punctuation etc. [19:20:20] anyway, the solution i found should i want to do morpheme by morpheme literal translation is just to replace grammatical forms noun, verb, etc, with codes for suffixes, because in the end neither apertium nor the learner needs to know whether the suffix or prefix is a noun, verb, etc at all, just whether its an "-es" "-s" "-on" "-ino", etc. so all i really need is one paradigm in the generative L2 monolingual dict, that just spits [19:20:20] out the morpheme as is, or else a trasnlation if that's what we want. [19:20:44] what languages were you working on? [19:21:42] am going to work on the eo-en pair with darshak and jacob until I have a better notion [19:21:53] ah ok [19:22:50] GF doesn't have a resource grammar for esperanto, so if you wanted both directions, that would also have been difficult [19:22:59] we do have interlingua though :-D [19:22:59] then maybe try just creating a local version en1-eo2 just to see what it looks like, and if it's convincing then work on a es1-en2 pair and an es2-en1 pair. [19:23:25] jejeje been over 20 years since i've looked at interlingua but could probably figure it out! [19:24:05] though basic english might even be better, hehehe [19:24:18] :P [19:24:38] just choosing esperanto because its a small wordset [19:24:38] if you were still interested, I could do a small version of en-es and es-en, and give the code to you [19:25:08] wow, that would be amazing! [19:25:28] you need to install a bunch of haskell stuff first to get it to work :D [19:25:57] apt-get install unicorns [19:26:03] the example I pasted last week, I was in a café in stockholm and did it in 15 minutes or so :P [19:26:20] spectie: hehe :P [19:26:27] more like cabal install unicorns! [19:26:48] i am headed up tomorrow to montpelier to comb Bernard's brain about apertium, but i will definitely continue to investigate gf if you think that's a viable option. [19:27:47] yeah, I think it's pretty easy to do just that, language A's syntax and language B's lexicon [19:28:02] but if you wanted to have morpheme by morpheme, then it's not that easy [19:28:32] haha I'll test something [19:28:33] a sec [19:29:06] in the end, presentation of the data to learners neds to be in a format they can use. am reading about Lexicographic information cost right now. [19:29:11] needs [19:31:21] *** Joins: myoriel (~Myoriel@aftr-88-217-181-187.dynamic.mnet-online.de) [19:31:26] oops, gotta run, baby is screaming! [19:31:30] bbiab [19:32:38] see you! [19:33:18] I was thinking if gf morphological constructors could be used with words of the other language, but that's probably pointless :P [19:33:37] here's an example of using swedish morphological constructors with spanish words http://pastebin.com/khWgn25F [19:34:58] casornas! :D [19:38:00] that was particularly well-behaving pair, because swedish verbs take indicative present, which ends in r [19:38:24] actually finnish constructors complain if words don't end like finnish words :P [19:38:40] like if they end in -ough [19:38:40] ? [19:39:07] yeah :-D [19:39:09] for instance [19:39:35] > cc mkV "enough" [19:39:36] Internal error in Compute.ConcreteNew: Applying Predef.tk: expected infinitive, found enough [19:39:47] for nouns it works though [19:40:17] :) [22:11:33] *** Quits: jmvanel (~jmvanel@18.111.115.78.rev.sfr.net) (Ping timeout: 240 seconds) [22:11:46] *** Joins: jmv (~jmvanel@247.212.139.88.rev.sfr.net) [22:16:42] *** Joins: jmv_ (~jmvanel@74.184.139.88.rev.sfr.net) [22:17:19] *** Quits: jmv (~jmvanel@247.212.139.88.rev.sfr.net) (Ping timeout: 256 seconds) [23:07:53] inariksit: thanks for looking at that! [23:07:56] if mixed morpholgy is to be used in two languages with same (in this case latin) alphabet, i think there needs to be some sort of marker (like the * in apertium) to show it otherwise it'll be very confusing. [23:09:27] i think i might try and find a way to color code the text or italicize or underline so its obvious where one language begins and another ends, otherwise with words like casornas only a person fluent in both will know whats going on... [23:11:52] thanks to you spectie for your input, too, it's great. [23:12:39] God natt!