[03:20:32] *** Parts: relrod (~relrod@redhat/relrod) () [05:14:30] *** Joins: drbean_ (~drbean@124.219.83.42) [05:42:13] *** Joins: proteus-guy (~proteus-g@2405:9800:b408:bc31:79c0:9e5a:4a4f:9e8e) [05:44:49] *** Quits: proteusguy (~proteus-g@49.228.115.2) (Ping timeout: 246 seconds) [06:08:35] *** Quits: Flammie (~flammie@sandbox.multilingua.uni-hamburg.de) (Ping timeout: 240 seconds) [06:41:06] *** Joins: Flammie (~flammie@sandbox.multilingua.uni-hamburg.de) [09:38:14] *** Joins: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) [09:39:10] morning! [09:43:19] all these CN -> NP -> CN functions are terrible when some of your determiners and modifiers come in the front and some in the end [09:45:10] \o/ [09:45:22] good morning! [09:47:01] *** Quits: annaaerinteenbot (~annaaerin@cse-principia.cse.chalmers.se) (Ping timeout: 255 seconds) [09:57:57] now I think I can say "my big glass of red wine on the hill", but combining that with ApposCN would be ... okay not very grammatical in english either [09:58:14] whenever I say that GF is so pretty, just remind me of this ^^ [10:04:03] *** Joins: karelisnotanna (~karelisno@2001:67c:2198:31:1db0:dde1:5007:1cb6) [10:08:45] *** Quits: karelisnotanna (~karelisno@2001:67c:2198:31:1db0:dde1:5007:1cb6) (Remote host closed the connection) [10:09:28] *** Joins: karelisnotanna (~karelisno@2001:67c:2198:31:1db0:dde1:5007:1cb6) [10:12:53] *** Quits: karelisnotanna (~karelisno@2001:67c:2198:31:1db0:dde1:5007:1cb6) (Remote host closed the connection) [10:30:36] *** Joins: annaaerinteenbot (~annaaerin@cse-principia.cse.chalmers.se) [10:33:30] do we have a new bot! [10:41:04] *** Quits: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) (Ping timeout: 260 seconds) [11:10:35] it seems to be having trouble with its vocalizations, poor thing [11:10:43] it is like a little baby trying to speak [11:11:00] i will add debugging to line 167. [11:11:34] *** Joins: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) [11:18:55] Lang> l -bind PhrUtt NoPConj (UttQS (UseQCl (TTAnt TPres ASimul) PPos (QuestIAdv why_IAdv (PredVP (DetCN (DetQuant (PossPron i_Pron) NumSg) (AdvCN (PartNP (AdjCN (PositA big_A) (UseN tail_N)) (MassNP (AdjCN (PositA red_A) (UseN wine_N)))) (PrepNP on_Prep (DetCN (DetQuant DefArt NumSg) (UseN hill_N))))) (ComplSlash (SlashV2a kill_V2) (DetCN (DetQuant DefArt NumSg) (UseN bird_N))))))) NoVoc [11:18:55] why does my big tail of red wine on the hill kill the bird [11:18:56] zergatik muinoaren gainean nire buztan handiak ardo gorri txoria hiltzen du [11:19:01] soo many fields in my CNs [11:19:38] but now I can handle these very important sentences \o/ [11:23:05] *** Quits: annaaerinteenbot (~annaaerin@cse-principia.cse.chalmers.se) (Ping timeout: 240 seconds) [11:23:26] *** Joins: lillann_ (53b95d0c@gateway/web/freenode/ip.83.185.93.12) [11:24:18] *** Quits: lillann_ (53b95d0c@gateway/web/freenode/ip.83.185.93.12) (Client Quit) [11:24:21] Lang> gr PhrUtt NoPConj (UttNP (DetCN (DetQuant DefArt NumSg) (SentCN (UseN ?) (EmbedQS (UseQCl (TTAnt TPres ASimul) PPos ?))))) NoVoc | l [11:24:22] the fingernail who becomes longer than Paris [11:24:41] this looks confusingly correct, but the "who" is actually a question-who, not relative [11:24:47] like "the question why she sleeps" [11:28:22] (daherb: we can use all these examples as motivation in our paper why evaluating RGL by using gr is silly) [11:28:39] (and to be sure we have enough material, let me generate 1000 more sentences and read all of them carefully ^_^) [11:41:22] *** Quits: jbalint (~jbalint@unaffiliated/jbalint) (Ping timeout: 260 seconds) [12:19:03] *** Joins: jbalint (~jbalint@24-177-124-81.dhcp.mdsn.wi.charter.com) [12:19:03] *** Quits: jbalint (~jbalint@24-177-124-81.dhcp.mdsn.wi.charter.com) (Changing host) [12:19:03] *** Joins: jbalint (~jbalint@unaffiliated/jbalint) [12:20:55] sounds like a job for recaptcha ... once they're done reading all the house numbers in the world, they can start distinguishing grammatical sentences lol [14:50:22] venicecat: ooh now I see how your name is written, I thought all the time it was Charlie! [15:00:04] *** Quits: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) (Ping timeout: 260 seconds) [15:13:51] *** Joins: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) [15:14:15] hahaha yar! [15:14:24] It's the english version of my mandarin name [15:14:46] hihi [15:14:57] and also the official spelling [16:13:16] *** Quits: proteus-guy (~proteus-g@2405:9800:b408:bc31:79c0:9e5a:4a4f:9e8e) (Ping timeout: 276 seconds) [16:19:59] *** Quits: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) (Ping timeout: 260 seconds) [16:24:50] *** Joins: proteus-guy (~proteus-g@2405:9800:b408:bc31:a1d5:650c:84bb:1f05) [16:44:33] *** Joins: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) [17:31:09] *** Quits: venicecat (05b30941@gateway/web/freenode/ip.5.179.9.65) (Ping timeout: 260 seconds) [17:49:19] *** Joins: venicecat (c30dd25a@gateway/web/freenode/ip.195.13.210.90) [17:59:10] we party with our laptops at the teahouse pavilion in the park by the river a little south of the university. there are tables here, if you want to keep working. [17:59:58] *** Joins: JuanDaugherty (~juan@98.4.124.117) [18:06:36] greeting [18:20:10] *** Joins: venicecat_ (c30dd25a@gateway/web/freenode/ip.195.13.210.90) [18:21:54] *** Quits: venicecat (c30dd25a@gateway/web/freenode/ip.195.13.210.90) (Ping timeout: 260 seconds) [18:51:59] *** Joins: venicecat (c30dd25a@gateway/web/freenode/ip.195.13.210.90) [18:52:49] *** Quits: venicecat_ (c30dd25a@gateway/web/freenode/ip.195.13.210.90) (Ping timeout: 260 seconds) [19:24:19] *** Quits: venicecat (c30dd25a@gateway/web/freenode/ip.195.13.210.90) (Ping timeout: 260 seconds) [20:49:57] *** Quits: JuanDaugherty (~juan@98.4.124.117) (Quit: Hibernate, reboot, exeunt, etc.) [20:58:00] why do the parts-of-speech labels in the syntax trees produced by the wide coverage demo look so different from the RGL? [20:58:37] I see DetCN and UseN and ConjNP instead of mkCN the_Det etc [21:39:54] also: how do i construct: "the pleasure of eating bread" [21:40:43] the_Det (mkN2 (mkN "pleasure")) ... (mkVP (mkV2 "eat") (mkNP (mkN "bread"))) [21:50:07] mkCN and the_Det are API functions, whereas DetCN and UseN and ConjNP are abstract syntax functions of the RGL [22:20:01] ah, i have not spent enough time with the internals. [22:20:13] have you any advice about the achieving of the gerunding? [22:20:55] progressiveVP : VP -> VP -- be sleeping [22:21:26] now how do i shoehorn that into a Noun? [22:26:59] maybe the answer is: i don't [22:47:48] not into a lexical category, but it can easily become a subject or an object via S [22:48:23] Constructors.gf: mkSC : VP -> SC -- to sleep --: [22:48:23] Constructors.gf: mkCl : SC -> VP -> Cl -- that she sleeps is good --: [22:48:32] that's an example how it could become a subject [22:48:45] let's see, an object for a N2 [22:49:13] "i remember the pleasure of eating madeleines" [22:50:11] okay, there doesn't seem to be an easy API way to go from SC to a NP [22:50:13] i_NP ... mkV("remember") ... N2 of ... progressiveVP eating madeleines [22:51:59] then i shall blaze my own trail through the internals! [22:52:08] yeah :-P [22:52:12] "its skin is too thick! i shall have to cut it open ... from the inside!" [22:52:23] or you can extract the s from the SC [22:52:57] and start from the bottom with mkNP (mkN (mkSC (progressiveVP blah)).s) [22:53:23] oh my [22:59:30] i will try [23:12:27] yay it worked [23:12:28] oper VP2N : VP -> N = \vp -> mkN (vp.s2 ! (RE.agrP3 RE.Sg)); [23:12:33] \o/ [23:12:47] you did it! [23:13:36] does that look OK? [23:14:03] awesome, I've been wanting gerunds too [23:14:19] yeah, that's pretty much the way to go when there's no API function for it [23:14:42] if possible, it's better to use categories with less params [23:15:05] because if someday someone changes the internals of the english grammar, including names agrP3 and Sg, then your code will break [23:15:20] what's the better way of setting up an Agr? [23:15:31] (asks the contracts guy) [23:15:52] whereas if you had made it into SC, you could just access the s field [23:15:53] CommonX.gf: SC = {s : Str} ; [23:16:00] i try [23:16:20] VP : Type = { [23:16:20] s : VerbForms ; [23:16:36] and [23:16:36] VerbForms : Type = [23:16:36] Tense => Anteriority => CPolarity => Order => Agr => [23:16:37] {aux, adv, fin, inf : Str} ; [23:16:41] so that's a lot of moving parts [23:17:23] the mkSC gives me an extraneous "to be" [23:17:32] i need to decopulate [23:17:34] oh okay [23:18:14] i could do a hack to case out the "to be", but is that really better? [23:19:14] heh, it's still simpler than figuring out what of the dozens of fields of VP to access and with which parameters, but sure it's ugly and prone to breaking again [23:19:33] if someone changes the mkSC function so that it doesn't add the "to be" [23:19:45] so you can do whatever feels to you less ugly :-P [23:19:57] if the Right Way is to submit a pull request against ExtraEng i am willing to go diving in the internals [23:20:00] the night is still young [23:20:30] yep! ^^ [23:20:34] for the Right Way [23:21:07] if you're doing it as a part of RGL, then it's fine to use vp.s2 ! (RE.agrP3 RE.Sg)), because it's at least internally consistent [23:21:24] if someone goes and changes agrP3 or whatever, at least they have to change it everywhere in the RGL [23:37:20] going to sleep now, good night! [23:37:34] so long, and thanks for all the fish!