[10:38:58] I realised that the difference between RelCN and RelNP is this restrictive vs. nonrestrictive relative clauses [10:40:18] *** Joins: laurettem (29382389@41.56.35.137) [10:40:24] hello laurettem! [10:40:43] Hi Inari! :) [10:41:49] how's it going? [10:41:59] I'm just back to work after summer holidays and working on somali RG [10:42:04] Well, thanks, and you? [10:42:11] How's that coming along? [10:43:22] had some tricky bits to implement, but I've been just persistent in trying to make it as small as possible :-D it's nice when working on something that isn't yet complete, so if I want to change a thing, I don't need to tear apart everything else [10:44:25] I feel like there's some bloggable content in here about how to combine parameters in imaginative ways, but I'll need to figure it out properly first :-D [10:45:45] Yes, I can imagine. I'm working on Zulu myself, and finding it nice and tricky at times. [10:45:55] oh and I met bruno in france! here's some pictures ^^ https://s.put.re/s2jsxjuV.37.jpeg and https://s.put.re/HdSdvekr.47.jpeg [10:46:07] how well do you know zulu yourself? [10:46:29] That looks like fun! [10:46:31] Not very well. [10:46:50] I have a decent book and now I've made contact with a linguist, which helps a lot. [10:47:38] yeah definitely! asking questions is important [10:47:51] I've got 3 different somali books, from 50s, 80s and 2010s [10:47:57] they all say a bit different things [10:48:10] I find it a challenge to balance staying close to the RG abstract syntax vs implementing a grammar that feels natural to linguists. [10:48:19] oh yeah definitely [10:48:20] Or at least which fits with the existing literature on the language. [10:48:29] hence imaginative parameter combinations :-D [10:48:49] like I spent a long time deciding that I put reflexive as an "agreement" and passive as "preposition" [10:49:12] Haha, yes, I've been dabbling with those myself. Will be interested to see any blog posts you write on it. [10:49:23] (still not 100% sure I keep it that way, but so far it seems to be producing sentences that fit the example sentences in my grammar books) [10:49:50] you should blog yourself too! :-D [10:50:08] I might. :P [10:50:18] What language family is Somali? [10:50:30] afro-asiatic, subgroup cushitic [10:51:16] Ah. Because I also try to keep in mind that I might want to bootstrap to the other southern Bantu languages, which is a bit paralysing... [10:51:38] yeah I can imagine [10:52:06] I think it's easier to just do one RG so you know it well, and then if you feel like there's something to generalise from, do it afterwards [10:52:51] I've got some ideas from somali that I could try on basque to optimise a bit, but there's absolutely nothing that could work as a functor [10:52:55] I guess that's the only way, really, especially if the comparative linguistics hasn't been done in enough detail to know how it might generalise. [10:53:02] yeah [10:53:42] I was actually hoping to run into you here for a different reason... [10:53:47] sure! [10:53:52] I have a question about modules and inheritance. [10:54:24] Is there something like parameterised inheritance? Don't know if I'm even thinking of it correctly. [10:55:03] in a way that incomplete concrete / resource / interface / all that stuff doesn't cover? [10:55:03] Basically, I have no trouble "open"ing parameterised modules, but I would like to inherit an incomplete resource grammar as well. [10:55:22] ah right, to inherit an incomplete to another incomplete [10:55:47] I think we tried it with ExtendRomance and it didn't work [10:55:47] Maybe it does, I'm not sure. Let me try to describe what I want to do... [10:55:51] okay [10:57:07] Ah, rats. Well, I have a Symptoms.gf module that extends Time.gf. I want to create more than one version of English, with slightly different surface forms to be used for displaying on screens and sending to TTS etc. [10:57:49] So, I have "incomplete concrete TimeEnx", which makes TimeEng.gf and TimeEns.gf etc. [10:59:05] I think that should be fine, if you can extract all the things that make a difference in TimeEng/TimeEns into resource modules [10:59:31] if TimeEng and TimeEns are concrete, not incomplete concrete [10:59:45] Yes, that part works. Now, I also want "incomplete concrete SymptomsEnx" to make SymptomsEng.gf and SymptomsEns.gf. But, I need to inherit TimeXXX somehow, so that I end up extending TimeEng.gf for SymptomsEng.gf and TimeEns [10:59:58] whoop,s for SymptomsEns.gf [11:00:20] incomplete concrete SymptomsEnx of Symptoms = TimeEnx ** open ... [11:00:31] concrete SymptomsEng of Symptoms = SymptomsEnx with (TimeEnx = TimeEng), ... [11:00:51] But then, TimeEnx is not opened, its inherited. [11:01:54] mm, I don't think I've needed this before, I suggest you write to gf-dev! [11:02:07] now you have typed the question so the hard part is done ^^ [11:02:18] Haha, okay! :D [11:03:06] I'll do that right now. Thanks for the chat! [11:03:11] no problem! [11:30:03] *** Joins: drbean (~drbean@TC210-63-209-67.static.apol.com.tw) [11:45:28] Btw, Inari, now that I've typed out my question (and sent it, haha), it seems clear to me that I am, in fact, inheriting one incomplete to another incomplete like you asked. Oh well. [11:45:28] *** Quits: drbean (~drbean@TC210-63-209-67.static.apol.com.tw) (Ping timeout: 245 seconds) [11:48:32] *** Joins: drbean (~drbean@TC210-63-209-22.static.apol.com.tw) [11:53:43] *** Quits: drbean (~drbean@TC210-63-209-22.static.apol.com.tw) (Ping timeout: 244 seconds) [12:12:31] laurettem: hehe right! it was useful to type it out in any case, and maybe someone answers with suggestions how to do it in another way! [13:12:51] some gftest poetry: "I can be a salt that the small father of you that is eaten teaches" [14:47:02] *** Quits: laurettem (29382389@41.56.35.137) (Remote host closed the connection) [14:51:35] *** Joins: laurettem (29382389@41.56.35.137) [15:37:21] *** Quits: laurettem (29382389@41.56.35.137) (Remote host closed the connection)