Wikipedia talk:Gather/Gather RFC proposal - February 2016

Initial response

Seems like a desperate attempt to save something that isn't worth saving. Delete it, and nothing of value was lost. The Quixotic Potato (talk) 07:07, 19 February 2016 (UTC)

Want a "more collaborative relationship"? Maybe you should stop beating that dead horse and listen to the community for once. This proposal is an insult to the community. If you guys would respect the community then you wouldn't propose this.

You don't have to export anything (there isn't anything worth exporting), and working on something like an export feature would be a waste of time. Discussing potential product arcs is something that you should've done a long time ago. Exploring potential solutions to problems with the feature isn't necessary, we already have a solution that fixes most problems: disabling Gather. We have consensus that this is the correct solution. Now it is time for you guys to drop the stick and back slowly away from the horse carcass.

You are unable to determine what a viable roadmap is and what isn't. WMF roadmaps are bullshit.

You are basically saying that you want to work for months on something that is useless, despite the fact that the community has explicitly told you that the community does not want you to.

Jimbo is not your boss. The WMF is not your boss. The community owns Wikipedia, because the community created it, and the WMF should serve and protect the community. Please go work on Phabricator tickets instead of wasting time on vanity projects. The Quixotic Potato (talk) 07:39, 19 February 2016 (UTC)

I don't get why your proposal is so complicated and doesn't follow the clear consensus at the RfC (yes, it isn't closed yet, but the outcome is quite clear): disable Gather now and only revive it after the steps set out in this "RfC proposal" (without a deadline) and a new RfC that explicitly gives the WMF permission to relaunch it on enwiki.

With "disable", I mean:

• Disallow the creation or editing of any further Gather collections, whether public or private
• Either set all existing collections to private (worst solution), or convert all existing collections to user subpages (for which you already have written a bot, as claimed by User:TNegrin (WMF) on 5 February) and delete the Gather collections, or simply delete all Gather collections.

According to the RfC proposal, you will only be able to set all pages to private at the end of March, so that's really not an acceptable response to the concerns in the "Disable Gather RfC". Please implement solution two (convert to user pages and delete) or three (delete) instead.

As for "More than twenty people from the Reading team participated in the discussion and collaboration that led to this proposal. I am telling you this because I think it shows the level of interest that exists around this feature and the subsequent community reaction. Personally, I’m proud of my team for being able to agree on the proposal showing the concerns they have for the community, the project, and collaborative decision making in general.", that's PR bullshit. For almost a year, there was zero interest in this feature at the WMF, then we got a so-called "community liaison" trying to avoid the RfC and the inevitable outcome by stonewalling us with some bare-faced lies, and when that didn't work we get this last-moment effort. The only reason many people at the WMF react now is because of "the subsequent community reaction" only. You may be proud of your team, but their "reaction" only shows the typical WMF panic reaction and desperate attempts to salvage anything possible from their unwanted, untested extension, just like we have seen countless times with previous similar things. At least you no longer can use Superprotect this time.

You claim as #2 of the "lessons learned" that you will "Understand and respect community culture, values and requests during all stages of development.", yet the outcome of this RfC (a community request) to disable Gather will be disregarded for at least three months (probably more, because who is going to decide if you have a "viable roadmap" after that period?). A great way to show that respect! Fram (talk) 08:09, 19 February 2016 (UTC)

This reponse is completely unacceptable. All but two participants at the RFC supported disabling, several of whom explicitly stated "immediately". Under your plan, it would take 4 months since the initial discussion behind the RfC began. This can only be seen as yet more stalling tactics, a continuation of the stonewalling provided by so-called community advocate Melamrawy. Not disabling Gather is dereliction of your duty, not only because of the very explicit opposition to Gather as it stands, but also because you are allowing a cesspool of BLP and NFCC violations, which cannot be moderated, to accumulate. Why will it take a month under your plan to make collections private? That's another month of opportunities for bad actors to create malicious, libellous Gather lists, with little chance of being spotted by an admin, and knowing that it cannot be deleted. I fail to see why you are proud of a team which lets a damaging "feature" linger and only starts caring when their pet project is threatened. You need to snap out of PR mode immediately. BethNaught (talk) 08:23, 19 February 2016 (UTC)

Lame

It is a bit of a lame response to an almost unanimous "disable now" to propose to hide the feature and then promise to maybe disable it if there are no new ideas in 60 days. Can't you just turn it off? Or even just set all collections to private now? Not disabling the feature even though it does not meet Risker's tests for a minimum viable product seems bad from a software engineering point of view. Not following community consensus on this is simply lame. I have an alternate suggestion:

1. Turn the feature off. In other words, Gather will be non-accessible on this wiki.
2. Once the feature has been turned off, start the clock on a 60 day countdown.
3. During this period, work on the following tasks:
   1. In consultation with previous users of the feature, we create a plan for export their collections.
   2. Discuss potential product arcs with the community (for example, collections, books, watchlists).
   3. Explore potential solutions to problems with the feature.
4. If we have a viable roadmap by the end of the 60 day countdown, we will keep the feature disabled.
5. Once we have a working product that satisfies Risker's basic viability standards, we consult the community about turning it on again.

In short, please remove all broken software from this wiki until you have fixed it. I am not interested in seeing a "roadmap" at this point while the broken software stays enabled. Please find out whether Gather can be fixed (it may be impossible because its design seems not to be wiki-compatible), then fix it, then come back. —Kusma (t·c) 10:10, 19 February 2016 (UTC)

Yes, when I read the first point of the proposal "1. Make the feature private", I assumed people at the village pump where overreacting to a very reasonable plan of action. The rest of this proposal then goes on to completely undermine that good first point... Why does even making the collections private, an action that would alleviate the most pressing issues while causing minimal disruption for the few people using this feature, be delayed for at least two more months? If you really need four more months to decide what to do/fix it after that, I'm willing to give it to you. (I'm also really disappointed that having a proper rollback plan isn't a requirement for a rollout of any new feature, but here we are...) —Ruud 12:40, 19 February 2016 (UTC)

Recent edits

@TNegrin (WMF): you have made two edits to this proposal. Are you unaware of the fact that this proposal has failed? The Quixotic Potato (talk) 17:20, 19 February 2016 (UTC)

Err not really enough debate to decide that yet.©Geni (talk) 20:10, 19 February 2016 (UTC)

Two announcements, one RFC and this proposal. Too much time has been spent talking about this. The Quixotic Potato (talk) 21:08, 19 February 2016 (UTC)

TTP: time to private

Hi, I know this doesn't address all of the concerns mentioned, but I just wanted to clarify something that seems to be confusing about the proposal TNegrin (WMF) posted on behalf of our team. We are proposing making collections private as soon as we possibly can- we are not proposing a 2 month wait to do that. We will start with our next available sprint (2 week cycle) and I believe a patch is already ready to be merged into the code to make this possible--we just did not want to make promises with specific dates since external events outside our team can threaten any specific date we set. --Jkatz (WMF) (talk) 17:31, 19 February 2016 (UTC)

@Jkatz (WMF): Hi! Thanks for your message. Can you please tell TNegrin (WMF) that I've asked TNegrin a question (above)? Thanks! Do you agree with this proposal? Why don't you make that same promise without a specific date? You can simply write: "The WMF promises to make the collections private as soon as possible", right? I hope you do realize that this is not a solution, it is just part of the solution. Gather needs to be disabled. The Quixotic Potato (talk) 17:53, 19 February 2016 (UTC)

Phabricator ticket

@BethNaught: @Fram: @Kusma: @Ruud Koot:

There is currently a phabricator ticket that says: "Disable Gather's public facing features".

This seems to be an attempt to trick us. The RFC clearly states that Gather should be disabled. The WMF is trying to hide it, so that they can pretend that they've listened to the community. The Quixotic Potato (talk) 18:25, 19 February 2016 (UTC)

A helpful bot added a related patch set:

Allow Gather to run in private mode

https://gerrit.wikimedia.org/r/267332

This is so ridiculous its difficult to express it using words. I will try interpretive dance. The Quixotic Potato (talk) 18:40, 19 February 2016 (UTC)

It seems that Jdlrobson doesn't understand what an RFC is. The Quixotic Potato (talk) 19:09, 19 February 2016 (UTC)

I really doubt the WMF can meaningfully be said to be doing anything at the moment. Setting the thing to private is a reasonable way to begin a gentle shutdown or rework into a better product. Alternatively its a neat attempt to get around the possible issue that the only people with authority to turn it off are a little busy right now.©Geni (talk) 20:16, 19 February 2016 (UTC)

Nah, it was a misunderstanding. First the task author made a mistake, because he thought that the proposal was the RFC, and when I fixed his mistake people reverted me so I had my first Phabricator edit war but now it seems that they understand. The Quixotic Potato (talk) 20:21, 19 February 2016 (UTC)

that's an imaginative interpretation of what happened on that phabricator ticket... Bawolff (talk) 00:15, 20 February 2016 (UTC)

Actually, that is what I believed at the time... but I was wrong. The Quixotic Potato (talk) 02:07, 20 February 2016 (UTC)

PLEASE CALM DOWN

I have been very critical of WMF engagement in the past, but this looks like a great improvement to me. I don't think we should be biting their heads off.

1. I agree that setting Collections private disables it for all practical purposes.
2. The RFC was explicitly about disabling it pending a further discussion of whether an improved version would be acceptable. I'd like to see it gone permanently, but there were a large number of !votes in the RFC that were at least ambiguous about possibly keeping an improved version. It is at least plausible there could be consensus for that.
3. The proposal here is to make Collections private one month after the RFC closes. The WMF is agreeing to what we asked for, we can show some good faith and agree that a month is not a completely unreasonable allowance to shift developers off of their current work, to code some changes, to test them, and to carefully manage the transition.
4. It establishes a two month time frame to sort out whether we might want to keep an improved version. That is perfectly compatible with the RFC. It also sets a default outcome of removing it completely. The WMF isn't fighting us. The WMF isn't cherry picking things to fix from the RFC and presuming that equals keeping it.

Am I missing something here? Are people really raging over the nuances of the initial "disablement", or allowing month to do the work? I'd be happy to see the WMF embrace this step forwards in WMF-Community engagement. Let's not make them regret it by acting unreasonable. I sympathize with all the venting of past frustrations, but it's unhelpful. The goal here is for the WMF and Community to work as partners to sort these things out. Alsee (talk) 10:36, 22 February 2016 (UTC)

So, you disagree, no need to shout about that. We aren't "raging", thank you, we are protesting. Raging is starting sections in all caps and the like. Let's see, what are you missing? Putting something private only or disabling it are not "nuances", they are quite a difference. The WMF gives no reason why they don't want to simply follow the RfC but instead propose this, which somewhat but not completely follows the RfC at first sight. They have declared on 5 February that they have a bot that can convert all Gather collections to user subpages, but when confronted with that claim, they now keep mum (just like about every other promise they made, which means that we have no reason to believe them here either), and come with this much weaker and slower proposal instead. They also don't indicate if they will still allow the creation of further private pages or not.

The RfC was not about "disabling it pending a further discussion of whether an improved version would be acceptable." it was about disabling it now, no discussion, but leaving explicitly the possibility of having a new RfC to re-enable it after the normal process to create such an extension would be had (i.e. discussion with the community about all aspects, creation and testing on a dedicated site, more testing, more testing, and only if they have tackled most issues, bring it back to enwiki to have an RfC to see whether we want to implement it here). No disabling only pending a new discussion , but disabling without excluding the possibility of a new discussion if they get a) agreement that Gather would be a good idea and b) get a good set of requirements together and finally c) get a decent working beta version (not the pre-alpha we have here).

Needing a month to automatically make all collections private (one of the two default states already existing) is not reasonable.

It doesn't establish a two-month time frame, the two months only start after the current collections are set to private, which will take a month, only starting from the 29th February, not even now. So that's at least a three-month timeframe, after which they will present their conclusions or proposals (perhaps, if they haven't forgotten this timeline by then), after which presumable a new RfC will need to be had to get the conclusion we had this time around (i.e. to disable that), but then not based on what we actually have but on some nebulous new promises by the WMF which we should know by now are almost never trustworthy.

To be fair, the proposal doesn't say that they will present their proposals and we will decide if it can be disabled or not, it states " If we do not have a viable roadmap by the end of the 60 day countdown, we will disable the feature." without any indication of who will be the judge of that. Based on the performance record so far, it certainly shouldn't be the WMF, no matter if they put two people or twenty people on this.

In summary, the "proposal" is not an adequate response to the RfC, it is a stalling process designed to make it much much harder to ever get the extension disabled here. It is the continuation of the tactic started by the so-called Community Liaison, completely absent from this discussion since his or her false promises became apparent to everyone involved, and now coated in some oh-so-reasonable proposal, designed to show their interest in "a more collaborative relationship" by coming with a one-sided proposal without community input, but reached by a team of more than twenty people in some untrackable off-wiki discussion, and resulting in a phabricator task that was indicated to be "requested at the community RfC" until it was pointed out repeatedly that that was obviously not the case, and that this task (disabling things) was an internal WMF decision.

By the way, it has been more than a month since the copyright issue with the fair use images was noted; have you seen any change so far? It was oh os urgent, but the WMF has done much discussing and very little action. No simple "hide images until we have a good solution", but keep everything in its problematic state until we have some solution eventually or until people have forgotten about this. Thanks, but I have zero trust in WMF proposed solutions and tactics, and no half-baked proposal will get my approval. That's not raging and shouting, that realism and experience. Feel free to disagree with my rejection and to accept the proposal, but don't try to silence people with a different point of view on this. Fram (talk) 11:34, 22 February 2016 (UTC)

Fram This isn't about winning or losing for me. It's not really about Gather. It's about finding a way for us to communicate and work together on something we both care about. I'm sorry about the past; I'm to blame for some of it, other people are too. But none of us can change it. So let's try to figure out how we can make it better in the future.

I've proposed a plan to address the obvious flaws in Gather, migrate existing users to it gracefully and discuss what the feature might look like in the the future. If you don't think the schedules above are reasonable, let's talk about it. Software schedules are notoriously tricky and the ones in the proposal are very conservative. People are working on other things, production issues come up, etc. I just want to make sure we can hit them.

In all this, there have been some very interesting ideas about shared watchlists and cross-wiki watchlists that could benefit both editors and readers. I'd really prefer to figure out how these will work rather than battle over a feature that we've all basically agreed can't continue as is. TNegrin (WMF) (talk) 15:06, 22 February 2016 (UTC)

I would be more willing to forget about the past if it wasn't so recent. This is about things that happened in the discussion leading to the RfC and things that happened during the RfC as much as it is about the original implementation of last year. You've proposed a plan, yes, but you still haven't e.g. indicated why the idea of migrating the collections to user sub pages, for which you claimed to have a bot ready on 5 February, during the RfC discussion, has been utterly abandoned and replaced with "keep Gather but make them private" instead, which has basically no advantages for said users at all.

Your plan addresses no flaws in Gather, your plan is simply a means to stop the disabling of Gather for unknown reasons (we know that e.g. VE was implemented and kept live when it was an utterly useless piece of crap because they needed to do so because of some Grant, but thanks to the great openness of the WMF we don't know if Gather has any similar baggage). I don't care about the schedules apart from when they are misrepresented / miscalculated, but it is obvious that even getting rid of the fair use images on pages where they don't belong is too much to ask from the WMF (not just in Gather, also in Related articles and other nonsense). But I have little interest in starting another RfC in four months time to decide whether you have presented a viable roadmap, and whether said roadmap is a good enough reason to keep it active here instead of e.g. testing it first on other wikis (which is also part of your proposal, but which doesn't seem to go beyond words). You are trying to reverse the burden of getting your botched extension active here: it has been active for months, we have now decided that it sucks badly and want no part of it; you now need to go back to the drawing boards and see if anything can be done to fix this, instead of begging here for a second chance because perhaps you will be able to fix it, if someone finds some time in their busy schedule.

Instead of proposing a half-baked proposal which creates some deadlines for your overworked team, you can just disable Gather, work on the ideas behind it (together with the community) without additional pressure, and then decide what the best solution for it would be. If Gather can then help in building that solution, fine: no one is asking you to do delete the code and burn it, just to disable it here. If Gather has no place in those solutions, also fine. But what you are trying to do now is putting the cart before the horse, making a desperate attempt to save Gather as if the extension is the important thing here, and not the tools wanted by readers and editors. That good ideas have been raised during this discussion (with many of the ideas about watchlists raised probably a decade ago already, but with the WMF ignoring them) has no bearing on this RfC and proposal as far as I can see.

So I agree with your conclusion, it would be much better if you (plural) figured out how these (watchlist ideas and the like) would and could work than to battle over a feature that isn't deployment ready at all now, if ever. But I don't see how your proposal fits into your own conclusion though. If you don't want to battle over the feature that was suspended and can't continue as is anyway, and if the RfC has closed with the clear recommendation to shut down Gather on enwiki completely, then why are you proposing something else? Fram (talk) 15:35, 22 February 2016 (UTC)

To put it briefly, we've had plently of time to talk about it. We talked about it in (example diff) April 2015 when the community made it crystal clear to you that we want fuck all to do with moderating it. We talked about it for a week before the RfC was launched - or we talked about it, Melamrawy blathered on, giving us corporate-speak and no substance. We talked about it for a month at the RfC. This has been talked about for a long time, and all you have proposed is that we spend another 2-3 months talking about it. Several people have supported the idea of Gather in the RfC, and even then demanded that it be disabled during the problem-fixing period. Is it not clear to you that in the matter of disabling Gather, our patience is now at an end'? BethNaught (talk) 17:51, 22 February 2016 (UTC)

TNegrin (WMF), maybe I'm making overly positive assumptions or maybe other people here are making overly negative assumptions. It would be helpful if you were extremely clear on the conditions required for a "viable roadmap" to keep Gather. Is it:

1. The WMF decides it has sufficiently addressed reasonable concerns raised in the RFC. (The RFC was a very strangely worded user bug report.)
2. A future RFC supports Gather with existing or proposed improvements.

If the answer is a clearly stated (2) then I am hopeful that hostility on this page will significantly diminish. If the answer is (1) then I'll apologize to my fellow editors & strike my previous post. Alsee (talk) 12:52, 23 February 2016 (UTC)

Hi Alsee The RFC has been closed and we will disable Gather next week. No further work is planned on this feature. I appreciate your feedback on this proposal. TNegrin (WMF) (talk) 18:14, 23 February 2016 (UTC)

Tagged as "failed" to avoid fruitless discussion

I have tagged the proposal as "failed" because of Wikipedia:Village pump (proposals)#WMF Reading team's response to the closing of the RFC, which is incompatible with this proposal and supersedes it. Fram (talk) 07:46, 23 February 2016 (UTC)