[03:48:38] *** Quits: ayberkt (~root@174.138.71.241) (Quit: leaving) [03:48:52] *** Joins: ayberkt (~root@174.138.71.241) [05:01:20] *** Quits: JuanDaugherty (~juan@98.4.124.117) (Quit: Ex-Chat) [05:22:26] I wonder if I should be using these functions instead of the overloading mkNP, etc Paradigm ones. I've only ever used the functions from the synopsis. [05:26:47] That table is linked from the RGL Synopsis page at "Internal abstract syntax: synopsis of internal abstract functions and their Universal Dependency labels", under the Other Relevant Documents heading. [05:54:02] *** Quits: proteusguy (~proteus-g@2405:9800:b408:bc31:6197:ea96:b761:dbe5) (Remote host closed the connection) [07:47:16] *** Joins: proteusguy (~proteus-g@180.183.115.189) [10:10:29] *** Joins: venicecat (caa61422@gateway/web/freenode/ip.202.166.20.34) [10:10:54] greetings [10:11:02] i have stable internet [11:16:24] woo! [11:16:33] are you back in singapore? :) [11:17:14] drbean: if you are doing application grammars, it's recommended to stick to the API functions in SyntaxXxx and ParadigmsXxx, because they are guaranteed to stay the same [11:18:42] of course if you e.g. have your own fork of the RGL and know that you will keep it stable, then feel free to use the internal functions; or simply if you don't mind to recompile your code from time to time when the RGL updates [11:19:47] for instance, maybe someday someone actually refactors dutch and afrikaans into a functor, and then probably some of the lincats would change [11:19:51] and the names of params [11:21:11] sometimes I also just look at some tables in RGL and think "dammit it would be prettier if I just flipped the order of these two params in a table, then I could use partial application in many other functions", but the fact that I know that some people hardcode stuff like cn.s ! ResXxx.Foo ! ResXxx.Bar makes me unwilling to do that [11:21:31] and to be fair, it's not such a great use for my time, I should rather implement new functions in a language that doesn't yet have them :-P [11:22:16] ("some people" totally includes me, btw :-P sometimes the API just doesn't cover all that one wants to do) [11:34:46] *** Joins: Flammie (~flammie@sandbox.multilingua.uni-hamburg.de) [11:37:59] *** Quits: proteusguy (~proteus-g@180.183.115.189) (Remote host closed the connection) [11:56:46] yes inari i'm back in Singapore! \o/ [11:56:59] after many train rides, plane rides transits [11:57:38] How was tallin? [11:57:52] i would like to visit some day [12:00:58] so being the gf addict that i am now, i'm back to working on my code [12:02:42] so a question: if i write something like this: [12:03:09] certify = (mkCl (mkNP (mkDet this_Quant))(mkVP (mkVP (mkVP D.certify_V)(SyntaxEng.mkAdv D.that_AdA (SyntaxEng.mkAdv in_Prep (mkNP D.exchange_N)))) D.for_Adv)) [12:03:51] i thought i should be able to linearize this [12:04:02] But lol nothing happens when i did a gt | l [12:04:27] but i fixed all the errors [12:04:37] my foggy brain says [12:06:28] let me share the code https://gist.github.com/anonymous/a0a68442460942006a22a2d9f242e166#file-concrete-gf-L16 [12:09:19] try to do gt -cat=Cl | l [12:09:28] your startcat is Contract, and certify is an Action [12:09:33] gah sorry, -cat=Action [12:09:40] my brain is foggy too :-P [12:25:31] oooh that's a new command [12:25:52] i learned a new thing today yipppeee [12:26:37] oh my oh my it works! [12:26:47] what does that command means? [12:27:26] when do i use that versus the simple gt | l [14:19:24] gt -cat=Xxx generates all trees in category Xxx, whereas gt generates all trees in the startcat [14:19:43] and in your grammar, you have no lins that would produce a Contract [14:19:51] so gt cannot find anything [14:29:18] can I suggest improvements to your grammar? [15:08:05] *** Quits: venicecat (caa61422@gateway/web/freenode/ip.202.166.20.34) (Ping timeout: 260 seconds) [15:46:24] *** Joins: JuanDaugherty (~juan@98.4.124.117) [16:47:37] *** Quits: JuanDaugherty (~juan@98.4.124.117) (Quit: Ex-Chat) [16:47:58] *** Joins: JuanDaugherty (~juan@98.4.124.117) [17:22:35] *** Joins: proteusguy (~proteus-g@2405:9800:b408:bc31:28db:4234:4232:6949) [18:12:02] *** Joins: koo5 (~koom@79.98.72.195) [18:24:59] Can I remind that it is not a very good GitHub practice to commit directly on `master`? [18:25:10] Here is a StackExchange question on this: https://softwareengineering.stackexchange.com/questions/223400/when-should-i-stop-committing-to-master-on-new-projects [18:25:44] I hope I'm not being annoying as I mentioned this before [18:27:19] But I just made a new PR to `GF:master` and realized that half of my commits is just merging from upstream :D [18:27:59] oblivious to the fact that stable and buildable are the same thing until the first mark release [18:28:10] Using feature branches also encourages everyone to keep semantically coherent chunks of commits together, which is good [18:31:11] can I do something to this on my own? like now I'm working on estonian, so should I just make a branch called "estonian" or something and keep committing&pushing onto it? [18:31:47] I don't mind at all, I know we are not used to GitHub and it's just good that we get new members in the community who can help us [18:31:48] inariksit: yes, exactly! [18:32:29] inariksit: if you want the things you are implementing to go sooner into master, you can also make even more specific branches like `estonian-adjectives` [18:33:11] inariksit: most of the times, the sequences of commits are for specific features [18:34:37] I make a branch for each module in the Turkish RG, but it doesn't matter that much [18:35:16] You can also have an estonian branch you keep pulling from if your commits are mostly arbitrary [18:36:47] I'm a bit undecided what to do, I was doing Extensions because they were mostly missing, and I found out that I have to change the structure of VP significantly; and then there's the "someday we should make a finnic functor" (especially if keeleleek writes the votic grammar!) [18:37:09] or to make it easy, I can just continue with basque :-P I just got inspired to work on estonian because I was in estonia for the weekend ^^ [18:39:28] so does this look correct? https://github.com/GrammaticalFramework/GF/tree/estonian [18:58:57] inariksit: looks good [18:59:04] okay [18:59:20] and should kaarel make his pull requests to the master or to my branch? [18:59:27] At any point you can click on Compare and see all the changes you've made after you branched `master` [18:59:54] They are working only on the Estonian grammar, right? [19:00:20] so far yeah [19:01:47] Yes they can either make PRs to your `estonian` branch or you can add them as a collaborator to GF and then they can commit directly to your branch [19:03:20] * mbrock chimes in to say that long-lived branches away from master have problems of their own, and it's not immediately obvious which approach is better... [19:04:47] mbrock: yes that is definitely a righteous concern---sometimes merging long-lived branches is a pain [19:05:33] But many people working on resource grammars usually don't step on each other's feet at all [19:08:51] it could be a problem if the abstract changes and then someone keeps developing a deprecated abstract in their own branch I guess [19:09:03] but it doesn't really tend to change, aarne is big on backwards compatibility [19:09:10] just adding new things and keeping old still there [19:11:00] ayberkt: I am reading your question on the pull request [19:11:21] (I'll add an answer to the PR too, just want to ask some questions here first) [19:12:22] as for gerund, there are actual abstract syntax functions that mention gerund in the name, so that's where you mostly have to think which one to use [19:13:02] if you have 3 ways to make a gerund, then for the RGL purposes you should just choose one that is correct, maybe silly or archaic, or overly literary, but in a sense "correct", in most cases [19:13:31] *** Quits: JuanDaugherty (~juan@98.4.124.117) (Quit: Ex-Chat) [19:13:35] and then you can add the two other gerunds in ExtraTur, with any name you want, like mekGerund : VP -> Whatever [19:14:13] or even just have opers in ResTur, mekGerund, meGerund, isGerund with appropriate types [19:14:28] and then decide which one of them to use in the various functions that may require a gerund [19:15:01] so if EmbedVP needs a gerund, that's cool because you can just pick one that works for embedded VPs [19:16:09] then you have in Extend (to replace Extensions) GerundCN, GerundNP and GerundAdv, and you can pick the same or another [20:15:10] *** Quits: proteusguy (~proteus-g@2405:9800:b408:bc31:28db:4234:4232:6949) (Ping timeout: 240 seconds) [20:27:37] *** Joins: proteusguy (~proteus-g@2405:9800:bc10:1ca:28db:4234:4232:6949) [21:03:27] *** Quits: koo5 (~koom@79.98.72.195) (Remote host closed the connection) [22:25:41] *** Joins: koo5 (~koom@79.98.72.195) [22:35:59] *** Quits: koo5 (~koom@79.98.72.195) (Ping timeout: 248 seconds)