Wiktionary talk:Votes/2009-03/Transwikis from other Wiktionaries

simple

Should simple be distinguished from other "foreign" wikts in this vote proposal? E.g.:

Either:

1. instructing importers to use this feature only for entries that will be re-formatted immediately after transwiki, and deleting such entries if they are not re-formatted immediately (since they can always be re-transwiki'd if need be), or
2. instructing this only for non-simple.wikt entries, but leaving it to each importer's discretion for simple.wikt entries, or
3. not instructing this at all (i.e., leaving it to importers' discretion).

For entries, either:

1. using the Transwiki: namespace as a "staging ground" (like with transwikis from Wikipedia), or
2. importing them directly into the main namespace (since they'll be reformatted immediately), or
3. importing them directly into the main namespace (since they're from simple.wikt or will be formatted immediately ), or
4. not specifying this (i.e., leaving it to importers' discretion).

...or something like that.—msh210℠ 19:34, 11 March 2009 (UTC)

I don't know. Personally, I would want to treat Simple the same way as the rest, because their entries are formatted so differently from ours (e.g., using POSes as L2 headers, because they only include English words) that I wouldn't want a special laid-back attitude about imports from them. But I see how other editors could feel differently (after all, anyone can fix up an import from simple, but it takes special knowledge to translate and fix up an import from anywhere else), so it could make sense to incorporate those options into the vote. —Ruakh_TALK 20:54, 11 March 2009 (UTC)

If its technically possible to enable only transwikis from simple, that should be an option. Another possible option would to be to allow transwikiing entries from the base language of a foreign language wiki. (i.e., only French words could be transwikied from fr, only German entries from de, etc. Trying to deal with translations of translations is likely to get very tricky. Carolina wren 22:40, 11 March 2009 (UTC)

It is technically possible to transwiki only from simple. The other suggestion is not possible technically, but can, of course, be policy.—msh210℠ 22:48, 11 March 2009 (UTC)

Are you advocating one of those positions, or just saying they're logical? If the latter, then I agree, but my personal instinct is to avoid creating an option simply because it "should be an option" or is "another possible option". The options I created the vote-page with were those that:

• were my own stances.
• were sensible defaults. (no change; technical change but no new policy; etc.)
• were formalizations of stances people have taken in the discussion.

and just those already added up to six permutations. If no one's advocating a specific other option, then I don't think we should add it just because it might make sense for someone to advocate it. (Exception: if two camps have a strong disagreement, then it would make sense to add some compromise positions in the hopes of promoting consensus.)

But, as I said in the beer parlour, I'm not claiming ownership of this page; if you disagree, you are free to modify it.

—Ruakh_TALK 00:33, 12 March 2009 (UTC)

I definitely feel that a policy against incorporating translations of translations via transwiki is a good idea. Limiting us to transwikiing only from simple would be the extreme version of avoiding translation problems. Carolina wren 04:25, 12 March 2009 (UTC)

Well, that's not the same as the above. "[O]nly French words could be transwikied from fr, only German entries from de, etc." is not the same as "a policy against incorporating translations of translations via transwiki": the former would prevent me from using fr to get conjugations of German verbs, or de to get a bunch of illustrative quotations for a French word. I agree that if fr:neu defines German (deprecated template usage) neu as “neuf” (as it currently does), then an editor shouldn't turn around and create neu#German with the definition “nine”. But your above-proposed restriction seems way too extreme. —Ruakh_TALK 13:06, 12 March 2009 (UTC)

I would have to agree with Ruakh here. Imports should be only from the word of origin's Wiktionary. I.e. German words from de.wiktionary, French words from fr.wikt etc. The only time that an editor could import from multiple Wiktionaries is when the word is used in multiple languages, such as the example Ruakh brought up. Anyways, it shouldn't really matter if they are going to be reformatted to fit this Wiktionary's formatting rules here. Cheers, Razorflame 16:22, 12 March 2009 (UTC)

Actually, you're disagreeing with me. And that's not what my example was about. :-/   —Ruakh_TALK 20:36, 12 March 2009 (UTC)

Sorry, I was not very clear with my wording. What I meant was that you could pick any one of those Wiktionaries who has that word in their language as the one to transwiki from. This should only be done on things that have multiple languages of definitions, i.e the word exists both as a German and a French word, so therefore, you could pick between either one of those WIktionaries to import from, but you can only use one of them to import. I wasn't saying that you could import from both of them, because that would be pointless. Cheers, Razorflame 22:03, 12 March 2009 (UTC)

Don't apologize, I think it's me who was unclear. My points are:

1. Like Carolina wren said, we shouldn't import French pages to get their translations for German words (and so on). There are lots of reasons for this. I gave a fun example: fr:neu translates German (deprecated template usage) neu as “neuf”, but French (deprecated template usage) neuf means both “new” and “nine”, so if we blindly translated fr:neu, we could get a very wrong translation. But even in less obvious cases, I agree with Carolina wren that translations-of-translations are a bad idea: translations are always a bit wrong, and in a translation relay, these wrongnesses add up.
2. Nonetheless, I disagree with Carolina wren's conclusion. Even though a blind translation of fr:neu would be bad, there are plenty of reasons why it might make sense to import it: it might have a thorough and well-formatted declension, or it might have a good selection of illustrative quotations, or — heck — it might do a good job breaking down the various senses and translating each appropriately. (This breakdown may not work perfectly for English, but it could be a good starting-point.) In other words, the problem is only with a blind translation of fr:neu. If you don't know German, you have no business importing German words from any Wiktionary — even de.wikt; but if you know German, and fr:neu has useful features that our entry for (deprecated template usage) neu should incorporate, then by all means, you should feel free to import fr:neu and fix it up.

Also, it's not true that "you can only use one of them to import". It's possible to import multiple versions of an entry and merge their histories. (It's not a perfect system, but it works.)

—Ruakh_TALK 22:47, 12 March 2009 (UTC)

Rewritten.

I re-did the whole vote, and postponed its start date. The old version is at Wiktionary:Votes/2009-03/Transwikis from other Wiktionaries/Version 1. The problem with the old version was that it could not readily be modified to support new policy proposals, since each new proposal would split almost every option into two, generating near-exponential growth. The problem with the new version is that there are inter-dependencies between votes, so after the first round of votes we may well need a runoff to sort out the results. Since there are new policy proposals (see above), the problem with the old version seemed more serious. Any thoughts? —Ruakh_TALK 20:13, 13 March 2009 (UTC)

Nope, no new thoughts about this new and improved version. Kudos for the nice layout :P. I can't wait for this vote to start :P. Cheers, Razorflame 07:50, 14 March 2009 (UTC)

Immediacy

"Administrators (and other privileged importers) should not import pages from other Wiktionaries unless they will be fixed up immediately for use here." How soon is "immediately"? I assume "fixed up for use here" means "turned into a normal entry", so why not go the whole hog and propose that only complete entries can be added to WT, not transwikied stubs? Equinox ◑ 20:29, 13 March 2009 (UTC)

Re: "How soon is 'immediately'?": I've intentionally left that vague, but the idea is that the chain of events should go something like this:

• import one page → fix it

instead of anything like this:

• import a bunch of pages → fix some of them → do something else

Re: " […] why not […] propose that only complete entries can be added to WT, not transwikied stubs?": Sorry, I don't understand. We have to import the page before we can fix it; and it's a given that if we import a page into the Transwiki: namespace, then we fix it before we move it into any other namespace. Right?

—Ruakh_TALK 21:00, 13 March 2009 (UTC)

What I meant to say is: how necessary is the transwiki step, if we insist that the fix-up is done immediately? This rules out any large-scale automation of the process, anyway, so people might as well write up the entries in Notepad (or what not) and add them to the main dictionary namespace directly. Equinox ◑ 21:14, 13 March 2009 (UTC)

I'm sorry, I still don't understand what you're trying to say. It seems from your comments that you might be unaware of some relevant facts:

• It's possible to transwiki-import a page directly into the main dictionary namespace, without using the Transwiki: namespace as an intermediary.
• One of the proposed new policies is that we do so. (link)
• Transwiki-imports preserve the page history, so they make it straightforward to comply with the GFDL.

Also, since transwiki-importing a page is very quick and easy to do, whereas fixing it up afterward is hard and time-consuming, I don't think there's any real benefit to large-scale automation of the former.

—Ruakh_TALK 22:08, 13 March 2009 (UTC) rewritten 15:27, 14 March 2009 (UTC), when I re-read the original version and decided it sounded rude

I agree with Ruakh here. The import function is the best way to retain the GFDL compliance of any entry that we import here because it saves the editing history of an article, including when, who, and what was changed of every edit that was made to that specific article. Therefore, this is the best way to maintain compliance with the GFDL :). Cheers, Razorflame 03:54, 14 March 2009 (UTC)

Do we need any Simple-English pages?

I've never visited the Simple English Wiktionary more than very briefly, but is there actually any content from there which we don't have already? Didn't we already have all the entries currently in simple:wiktionary at least 2 years ago? I don't see the point of transwikis from there at all. --Jackofclubs 09:05, 14 March 2009 (UTC)

See Wiktionary:Beer parlour#Request for importer flag. Though, some of the examples there are obsolete, because Equinox (talk • contribs) has now created entries here. —Ruakh_TALK 15:32, 14 March 2009 (UTC)

I could probably find more where those examples came from, but I'm going to leave those out of this discussion as of what happened the last time I put them into a discussion :P. Cheers, Razorflame 05:49, 16 March 2009 (UTC)

What to import?

What sort of things would be expected to be imported? I've seen that e.g. lots of en.wikt entries have been imported into ar.wikt (see e.g. 4-4-4-4s). But AFAICT they haven't done much localizing with those entries. --Jackofclubs 09:10, 14 March 2009 (UTC)

Well, when I'm creating an entry for a Hebrew word, I'll often check to see if the Hebrew Wiktionary has an entry for it, and if it does, then I'll often use it as a basis for our entry — taking their niqqud (supplementary diacritics), their etymology, their sense breakdown, their example sentences, etc. Currently, I'll usually write a first pass of our entry, and then start to incorporate their information; but another approach would be to start by transwiki-importing their entry wholesale, and going from there.

I expect that not all editors will use this approach, but I think it should be available to those who want to try it.

—Ruakh_TALK 16:02, 14 March 2009 (UTC)

Last call for fixes!

The vote is set to start in about 28½ hours. If you have a chance between now and then, please re-review it and make sure I didn't miss anything. If any changes are still needed, please either make them, or postpone the start-time and leave a note for me to do so. Thanks! :-)   —Ruakh_TALK 19:37, 18 March 2009 (UTC)

Vote started

Unless there is clear anguish for me starting the vote, I am going ahead and starting this vote because it is past the date that it is supposed to be started. Cheers, Razorflame 00:40, 21 March 2009 (UTC)

Thanks! —Ruakh_TALK 00:49, 21 March 2009 (UTC)

Support votes not working

I voted support on two of the proposals but it has left the unformatted template there instead of expanding it to my name and a green tick. Can someone fix it? Equinox ◑ 00:51, 21 March 2009 (UTC)

Where's Robert when you need him?

"Oh, what's in a namespace?" currently has six votes in support of proposition 1, one in support of proposition 2, and five opposing both. (This may not last, but I'm writing as if it will.) Very likely (though not necessarily), everyone voting for either of the propositions would prefer to oppose the other proposition than to vote for it; so we (very likely) have six votes against proposition 1 to balance the six in its favor. It seems then that nothing wins this vote. Does that default to an win for those opposing both propositions?—msh210℠ 17:22, 14 April 2009 (UTC)

Honestly, I'm not even sure about how to handle "Thou shalt not cause languor": the new policy clearly doesn't have consensus (9–5–2 isn't even a two-thirds supermajority), but not having this policy has even less consensus (5-9-2 is barely a one-third submajority). Usually we retain the status quo ante bellum unless there's a consensus to discard it in favor of something else, but in a situation like this, was there a status-quo policy of "these transwikis are always forbidden" because the feature wasn't turned on here, or a status-quo policy of "these transwikis are always permitted" because we had no policies against them? (The latter is implicit in the votes' structure — the implication is that if a given restriction isn't proposed on this page, then we certainly won't have said restriction if we turn on this feature — but there's no reason we have to apply its logic to the restrictions that are proposed here.) Or do we just say there was no status quo, because the feature didn't even exist yet, and accept majority rule? (This last approach makes most sense to me, but it's slightly disconcerting to think that a simple majority can establish a status quo that then can't be changed without consensus.) —Ruakh_TALK 02:31, 15 April 2009 (UTC)

It seems to me that the support side is favoring the institution of a policy, while the oppose is favoring no new policy. It seems to me that in all cases of no-consensus, the result has to be "no new policy," which would allow languor. -Atelaes λάλει ἐμοί 05:43, 15 April 2009 (UTC)

Yeah, I suppose you're right. And I guess that conclusion is supported by the overwhelming consensi to request this feature whether or not any of these policies are adopted. —Ruakh_TALK 16:42, 16 April 2009 (UTC)

Note that if my analysis of "Oh, what's in a namespace?" carries the day (or if, for any other reason, we allow importation into, at least, the main (0th) namespace), DCDuring's "oppose" vote on "Thou shalt not cause languor" was explicitly only for the Transwiki namespace. (DAVilla may have meant this, too.) Thus, for the main namespace we have 9–4 (or 3 without DAVilla), which is probably enough. (Actually, if the vote is restricted to mainspace, Opiaterein seems to be voting in support, which boosts it to 10.) The decision would, I think, then be "thou shalt not cause languor in mainspace". What do you think?—msh210℠ 21:23, 17 April 2009 (UTC)