A method for proposing a better value for a specific feature for a specific language is needed

[HedvigS]

Hi, I'm a linguist who works on some of the CLLD-databases. I've been aware of CALS sometime but only now joined. I was wondering, users can only add entire languages right, not propose edits for datapoints? Is that true? Sorry, still getting used to how it all works.

I'd be keen on a model for the other CLLD-databases where there's a curator of each lg and users can propose edits, that can be approved or not.

[HedvigS]

And yes, I see that users can add comments but I meant something more streamlined like "propose other value for this feature for this language and add small comment on it and have curator/manager accept or reject". Pull requests basically.

[kaleissin]

This would mostly be relevant for nat(ural) lang(uage)s and there I basically try to mirror whatever's in WALS. With conlangs, there's generally a doting "parent" who knows best, right? Considering RMW Dixon's critique of WALS (Basic Linguistic Theory vol III, Appendix I, page 461 in the paperback) I can understand the desire to fix wrong datapoints in WALS =) I'd like to see more languages added, too, and more points for languages that interest me :) But: what is needed is a process for fixing the wrong points at WALS, and maybe a better process of quality assurance. On the other hand.. if we had a "pull request" for datapoints here, maybe we could use that there to push for changes? Hm. Feature request it is. stamped

Another thing to think about is that while conlangs at CALS all have a manager and may have editors, the natlangs copied from WALS don't, really. So we'd need a way to elect a curator/manager as well.

Btw, CALS is older than CLLD. It started as a reverse engineering effort of WALS pre-CLLD, including a lot of retyping texts by hand (thanks, some of the ppl with gold badges!). Natlang datapoints were only added after WALS got an export-function. I am planning to hook up to several CLLD resources though, especially Phoible and glottolog, at least at first.

Priorities-wise, being able to vote in new translation excercises, a working dictionary subsystem and a working phonology subsystem (with a hook to phoible!) is ahead in the queue. I do plan to have a system for adding new datapoints, with a voting system, and adding various missing values to existing datapoints. Currently, Egyptian and Hittite writing has no suitable datapoint for instance.

[HedvigS]

Hi. Thanks for response. I'm actually not that concerned with WALS, more the others that are still expanding. Have you seen Nordhoffs initiative for crowdsourcing WALS btw? It didn't get enough support I think, that's why it didn't progress (http://blog.cyberling.org/node/996)

I get that conlangs have doting parents to a higher degree, but with some of these other databases that are cropping up (CLLD and non-CLLD) there are often devoted coders or managers that take extra care of certain languages as well. Just look at APiCS for example, or the pronoun paradigm database etc. I can list more.

I'd be much interested in seeing how you tackle these issues and if it's possible to transfer that to other CLLD-databases. I don't have final call on these things for CLLD in anyway, but I'm keen to explore and forward new methods.

Perhaps this is when I should share this to you: https://docs.google.com/document/d/1qT5IgrLcETuZiA2ZH7kltdF_zwsg4IvxXQFaVqjpjUA/edit?usp=sharing

[HedvigS]

If I can get your email I'll add you to a mailing list for cross-linguistic databases.

[HedvigS]

Förresten kan vi prata svorsk istället om du vill :)

[kaleissin]

Hehe ja.