Latest comment: 7 years ago by Sänger in topic RfC: Disable Gather on enwiki

Add comments here!

Pau's comments[edit]

From [1]:

21:06:31 <cscott> I think Pau raised some good issues about the collections/page list feature on the mailing list today.
21:11:51 <cscott> pau said, eg: "I was wondering if collections would work cross-language (that is, based on Wikidata IDs). That would allow people to consume lists in different languages regardless of the language of the creator of those lists and would simplify the life of users participating in multiple wikipedias. On the other hand it would require to deal with articles missing in the local language (e.g., fallback to another language? ask users to translate i
21:32:56 <cscott> pau actually made a suggestion of that form as well:
21:33:02 <cscott> Considering that articles are living entities, it would be interesting to surface some updates about the content included in a collection. This is something that already happens in the watchlist, but I was thinking about something more focused on readers where I could view that an article was added to a list on interesting architects, or some piece of information was added there.

Can we have a copy of this conversation/of Pau's email? --Nemo 11:02, 22 January 2015 (UTC)Reply[reply]

I just sent a copy over. It's essentially what cscott pasted verbatim. Jkatz (WMF) (talk) 20:53, 4 February 2015 (UTC)Reply[reply]


In the draft schema, I only see things that an on-wiki list page can already do, except the "private" flag which you don't plan to use by default. If you don't need other data, I suggest to keep using on-wiki pages, to that you'll be compatible with Extension:Collection. If you need other stuff which can't be added there, please point it out and at any rate use another name for this thingy. --Nemo 11:10, 22 January 2015 (UTC)Reply[reply]

I believe we are prototyping this on wiki pages, but in the form of a json blob. Another name has been chosen, at least for the project: Gather. We might still call the feature Collections, since 99.9% of readers are not aware of the extension named collections. Jkatz (WMF) (talk) 20:53, 4 February 2015 (UTC)Reply[reply]

Watchlist wishlist[edit]

As per the discussion on Wikimedia-L this week, I'd like to know if this project is going to help fix some of the longstanding requests for watchlists (see Watchlist wishlist). It seems that many of the things that are asked for over there are related to what you're doing here. Personally - I would REALLY like to be able to group my watchlist so I can collect different kinds of articles that I watch into "folders" for easier management. Equally, I would like to be able to follow someone else's watchlist folder (if they chose to make it public). This seems to be a similar functionality to what you're looking at doing here, with the difference being that I'm thinking about it from an editor's needs in mind, while this project (as currently written) seems to be focused only on readers.

On a related note, can you tell me how the Gather extension, differs from the collecting tools in the Education extension, the Book extension, and the core feature of categories? This might be useful to add to some kind of FAQ. Wittylama (talk) 18:39, 4 February 2015 (UTC)Reply[reply]

Hi Wittylama. To your first point/concern, you are right. This should do a lot of what you are looking for, but is aimed more towards readers. The current idea is that you will have a watchlist, which is private, has recent changes, email notifications and other features. You can also have other lists, which are lists of articles with names, but do not have these editor-oriented features. To the extent that you would like more watchlists in order to bookmark articles, collect articles you have started,written, etc, this will help you. To the extent that you want to track recent changes, it will not. To the specific issue of multiple 'watchlists' with that specific functionality, I recommend supporting this proposed approach to this rfc.
To your second question, we have examined all of these and decided against building new functionality into these existing features for now. One big reason for this is that we want to experiment with different features and experiences as quickly as possible--this is facilitated by building something light-weight from scratch as opposed to modifying something that already exists and which people rely on on a daily basis. Regarding functionality and how it differs, it is most closely aligned with user-moderated books. One reason we are not simply porting that to mobile is to escape the mental model of a "book" and the limitations that imposes on the concept of collections. For one thing, it hinders people from creating lists like "Top 10 most influential philosophers" or "The pages I contributed to and am most proud of". All of the other features you mention are comprehensively driven by the community. Because each list is not exhaustive, there is a possibility of using popular lists to promote meaningful content. It's nice to know the full list of statisticians is there, but I might want to find a subset that have been picked out by a human for one reason or another. This is truly for different users, for a different use cases, and we are building it as an experiement, so we have different technical needs. Does that answer your questions, create new ones? Let me know. Jkatz (WMF) (talk) 20:42, 4 February 2015 (UTC)Reply[reply]
Thanks for the quick reply Jkatz (WMF). I feel similarly uneasy as others on this talkpage about the "starting from scratch" approach that's being taken with Gather, this may make it easier for you to iterate, but it also means that the WMF is increasing the number of somewhat-working extensions that eventually get abandoned when a new shiny thing comes along to play with... I'll take 1 fully functioning extension over 5 "experimental" extensions any day. I also wish to express my dismay that Yet Another Feature That Appeared Out Of Nowhere™ is being prioritised over features/improvements that have been requested frequently over a long period. That doesn't mean that Gather is a bad idea - but that the WMF only ever appears to work on "new" ideas that have originated "in house" rather than taking suggestions from the community on what should receive [some] prioritisation for new features.
To your first answer - From an editor's perspective, I do think the ability to create a personalised list (e.g. "articles I want to work on", articles I wrote) that can be shard - or even better, embedded - would be quite fun. Many (most?) experienced wikipedians have lists of article titles on their userpage that they wish to highlight. Equally many wiki-projects have lists of articles they wish to give attention to (as good, or also bad, quality). This tool could facilitate that kind of sharing. However, the fact that this tool doesn't intend to have any editor-specific features (like the watchlist does) reduces the functionality to merely sharing/highlighting [not that sharing is bad, just limited]. As for the phabricator and RfC links you gave me - as a non-technical user I have no idea what they mean.
To your second answer what you're describing sounds exactly like what the Special:Book tool does, just for mobile. A person manually selects a series of articles, as many as they want, puts them in a sequence they like, gives the collection a name, and presses save. English Wikipedia already has over 26 thousand user-created collections like this (obviously many of them would be people 'testing' the book tool etc. but the system works. Other than you're working for mobile only, what difference would there be in Gather from me creating a Book in that system and calling it "Top 10 most influential philosophers" or "The pages I contributed to and am most proud of" (your examples)? Wittylama (talk) 17:02, 5 February 2015 (UTC)Reply[reply]
These are very legitimate concerns, and if I were you I would probably be asking the exact same questions.
  1. On experimenting with new rather than fixing old. This is a balance every organization/movement has to achieve. I agree with you that the ground is littered with abandoned projects and as an organization, we are trying to get more serious about cleanup. If and when we decide that Gather is a winning feature, we will revisit how to best fold it in to existing features or fold existing features into it. Since we live on the internet and the paradigm is shifting constantly, I believe it is not enough to fix existing projects and maintain status quo--we always have to be experimenting with new features and new directions. Not every team, all the time, but that is what my team is tasked with right now.
  2. On building products that come from in-house rather than only building what editors ask for. This is also a balance. Again, most of our engineers are working on editing tools and improving our infrastructure. But the majority of our users (readers), the people whom all of us ultimately serve also have needs and these needs are not often voiced. So while some teams work on editing features and infrastructure, my team is experimenting for readers.
  3. As far as the Book tool goes, while you are correct that it does technically most of what we are looking for, it is built with a different use case in mind and this has technical implications. I spoke to one of the current maintainers of that code, C.Scott, and he thought it would be a mistake. Another employee, when asked if he thought the extension should be adapted for something similar said "only given Extension:Collections a cursory thought, it's pretty funky, JS-based, and geared towards pediapress rendering. ". Gotta run. Thanks and please keep questioning! Jkatz (WMF) (talk) 17:48, 6 February 2015 (UTC)Reply[reply]

STOP NOW[edit]

In case it's not clear, let's try a concise header. This proposal is nowhere near being designed, let alone implemented or deployed. gerrit:187716 is ridiculous. --Nemo 21:21, 4 February 2015 (UTC)Reply[reply]

Hi Nemo. It sounds like you are frustrated. I am sorry. Understanding that this is intended to be a prototype for rapid testing and iteration, can you provide some specific areas where you feel the design is lacking for the stage we are in? In doing so, please refrain from unhelpful insults like 'ridiculous' (perhaps you meant 'grossly insufficient')--my team is working hard and in good faith. Jkatz (WMF) (talk) 22:59, 4 February 2015 (UTC)Reply[reply]
I can't imagine a single reason why user collections do not solve the "user problems" listed on Gather.
We should not be overloading even more the terminology around collections/books. And more importantly, we should not have to maintain yet another set of lists of pages... Helder 03:32, 2 April 2015 (UTC)Reply[reply]
+1. Ill-conceived project. Stop now. Instead work on core architecture that serves all of wikipedia, and especially do not encourage non-ecyclopedic trivial content on a "mobile-only" fork of wikipedia, especially if it doesn't create wikitext pages subject to the universal "laws of physics" of the Wiki. 2601:0:9B40:D0:C1BA:669F:6641:BD1F 13:55, 12 May 2015 (UTC)Reply[reply]

Gather FAQ[edit]

Thanks @Wittylama: for suggesting the idea of an FAQ. Based on the discussion here and on other venues, I have put together a list of FAQ for Gather, feel free to add more questions to the list if needed, and headsup to @Nemo bis: as well :). Thanks--Melamrawy (WMF) (talk) 21:33, 10 March 2015 (UTC)Reply[reply]

Social Networking Entertainment Site[edit]

So... the idea for Gather is that readers, who don't edit, and don't log in, are supposed to start logging in so they can engage in Facebook-style social networking? So they can publish personal-blog-style random junk lists of their favorite bands? Am I understanding this project correctly?

The WMF sits for years ignoring useful improvements such as Watchlist_wishlist, and instead has this weird fetish for trying to turn Wikis into Facebook. Alsee (talk) 20:26, 30 March 2015 (UTC)Reply[reply]

Hi Alsee There is no specifics on how a feature "should" be used. It all depends on how users interact with it or even ignore it. As you mentioned already, something similar has been suggested years ago on Watchlist_wishlist, which is now being tested on mobile web beta, and will move with lessons learned and code and moderation rules, to desktop, as explained in the FAQ for Gather. You can test the feature here, here-- (please go to settings and mark beta options) and share your feedback. The FAQ also includes links on how the tested feature, while it is still limited to mobile, can still be use to overlap with desktop uses. Thank you --Melamrawy (WMF) (talk) 22:15, 30 March 2015 (UTC)Reply[reply]
That's a pattern, see MV, Flow and now this. They don't care about an encyclopaedia any more, they just care about click-counts and other rubbish. --♫ Sänger - Talk - superputsch must go 21:17, 30 March 2015 (UTC)Reply[reply]
Let me try again. Wikipedia is not a social networking site. We even have have a policy section EN:WP:NOTSOCIALNETWORK.
Wikipedians have their own user pages, but they should be used primarily to present information relevant to working on the encyclopedia. Limited biographical information is allowed, but user pages should not function as personal webpages or be repositories for large amounts of material that is irrelevant to collaborating on Wikipedia. If you are looking to make a personal webpage or blog or to post your résumé, please make use of one of the many free providers on the Internet or any hosting included with your Internet account. The focus of user pages should not be social networking, or amusement, but rather providing a foundation for effective collaboration.
We delete stuff like this. People don't come to Wikipedia to socialize or play. The reader-persona shows up to find useful information, and the editor-persona shows up to do work. That's why you're getting negative reactions here and on the Wikipedia Administrator Noticeboard discussion I saw. The WMF ignores useful stuff like Watchlist_wishlist for years, and instead makes a string of social-network-inspired projects that keep going over like a lead balloon with the Community. Alsee (talk) 22:12, 3 April 2015 (UTC)Reply[reply]
Hi Alsee, I guess the concept of making a combined list of articles isn't new, given that the books tool exist, and the Watchlist_wishlist part has already been covered in the FAQ, it is not ignored :) --Melamrawy (WMF) (talk) 23:33, 7 April 2015 (UTC)Reply[reply]
I guess it was ignored, until the communities rubbed the ignorant faces of the WMFers in this, and they were surprised to really have some kind of community wish existing, something no-one in SF usually cares about. So they just dumped it in the FAQ to pretend caring.
The WMF should stop all work on any new pet-projects by staffers (especially anti-policy facebookisation nonsense) until at least 80% of the existing community wishes are dealt with and are broadly accepted by the community. --♫ Sänger - Talk - superputsch must go 15:46, 8 April 2015 (UTC)Reply[reply]

Naming - "Favorites"[edit]

From what I heard, the word "collection" relates to wikibooks, "list" is used for many different things and is reserved in some languages. Has the word "Favorites" been proposed? From the start it implies that this feature is not neutral. --Yurik (talk) 04:25, 2 April 2015 (UTC)Reply[reply]

Hey Yurik did you check the renaming survey? :) The answers to the survey questions do include some interesting suggestions as well. --Melamrawy (WMF) (talk) 14:42, 3 April 2015 (UTC)Reply[reply]
I guess none of them edit on Wikibooks or know about Collection extension or List namespaces on some Wikipedias... Helder 02:06, 4 April 2015 (UTC)Reply[reply]
  • Yes, we saw the "renaming" survey. The option of "collections" should not have been presented, because collections already exist in the Mediawiki world. The survey should be redone without using the option for "collections". The options that were provided weren't very good, either; five minutes with a thesaurus would have come up with better options. Risker (talk) 15:01, 7 April 2015 (UTC)Reply[reply]
OK, I added a new section to the survey, to list some of the new suggestions.  :-) --Melamrawy (WMF) (talk) 23:41, 7 April 2015 (UTC)Reply[reply]

Collections that can be used just like the normal watchlist[edit]

Sorry if this is a repeat. I think the new Gather or collections feature is very useful, but it'd be even more awesome if you could use it like a watchlist, where it would show the recent changes to all the pages in that collection. This is something I've always wanted, which is why I made customWatchlists, a poor and hastedly put together attempt at the same idea. A pretty, native feature like this would be a dream come true! talk 20:30, 14 June 2015 (UTC)Reply[reply]

hi MusikAnimal, would you mind creating some watchlist-like collections and letting me know what do you think of this view? or this: Jkatz (WMF) (talk) 20:42, 14 June 2015 (UTC)Reply[reply]
@Jkatz (WMF): Yes yes yes! I didn't see a way to get to this through the collections interface? Anyway this is precisely what I've been looking for :) All we need is a friendlier desktop interface and you'll have our long-term editors mouth watering jaw dropped. I'm glad I didn't put much more effort into my silly user script. Thank you!! MusikAnimal talk 23:56, 14 June 2015 (UTC)Reply[reply]
@MusikAnimal: I am glad to hear that! We haven't built this aspect up to much or promoted, as Gather is designed for readers as a way of surfacing content to one another, but it is promising and exciting that you find it helpful. As you continue to use it, please let me know how you feel it could be improved to serve editor use cases as well. In particular, I am curious to hear your thoughts on how we could/should merge "collections" with watchlist. Do they end up being the same thing with different views or do we keep them mostly separated as now?Jkatz (WMF) (talk) 00:17, 15 June 2015 (UTC)Reply[reply]
@Jkatz (WMF): The approach you have now is the right direction, I think, where the two are merged and the watchlist is just treated as another collection, and selected by default. Collections/normal watchlist in the same view makes sense to me (this is how my user script works). Only thing is with desktop it'd be nice if we had all the features that we do with the normal watchlist. If you'd be able to pull that off I think few would object, as it'd resemble and work the way it always has, except you can also make your own custom watchlists through collections. In my opinion dumbing down the interface from what Special:Watchlist is now is probably a good idea too (make it more intuitive, the basic namespace tabs are a good idea), but the old timers will certainly throw a fit. So for desktop maybe focus on just combining the new collections into the old watchlist then work your way toward a redesign. MusikAnimal talk 18:55, 15 June 2015 (UTC)Reply[reply]
@MusikAnimal: Thanks for the input!. Yes, we are going to have to take baby steps! As you can see from the section below, we are putting development on hold while we figure out next steps, so this is great feedback on the next important hurdles is good to chew on in the interim.

@Jkatz (WMF): Is there an API for the new Gather feature? If not could we create one? As described above I had my own implementation of a similar feature, and it would take considerable time to port my custom watchlists to Gather. If there were an API endpoint I could write a script to do it for me. Thanks!! MusikAnimal talk 04:09, 21 November 2015 (UTC)Reply[reply]

Hey @MusikAnimal: , I apologize for the delayed response. I need to setup better notifications. This is heresy, but the best way to reach me is at my email jkatz [at] I am not aware of an API for this specific view, but the view was built using a series of API calls. Jdlrobson) made the page, but did so by stringing together a bunch of API ain't pretty.

For reference[edit]

--Atlasowa (talk) 23:26, 20 June 2015 (UTC)Reply[reply]

Gather violates English Wikipedia policies in an undetectable manner[edit]

Some Gather collections, for example this one consist entirely of non-free images, making it a non-free gallery which is not allowed by project policy. Even worse, the fact that non-free images are used contrary to policy is impossible to detect. For example, w:en:File:Asterixcover-20.jpg is one of the images used in the Gather gallery, but the "File usage" section does not mention that the file is used outside of the article namespace.

Can either images in Gather be disabled, or Gather not display non-free images, or Gather be disabled on the English Wikipedia for the moment? I don't think Gather is necessarily a bad idea to have (see my comments on w:en:WP:VPR), but I believe it must not violate local policy. An exception for Gather should be proposed before non-free images are displayed. Kusma (talk) 20:22, 20 January 2016 (UTC)Reply[reply]

Additionally, for free images, I am not sure that whether having the image attribution two clicks and possibly searching through the next page (if the image that is used isn't right at the top) is compatible with all of the image licenses we allow. This should be checked in detail. Kusma (talk) 21:16, 20 January 2016 (UTC)Reply[reply]
Since MV and superprotect we all know, that the WMF doesn't care that much about proper follow licenses, as long as it's a shiny piece of bling that's been released. --Grüße vom Sänger ♫(Reden) 05:24, 21 January 2016 (UTC)Reply[reply]
I don't think this attitude is very helpful to WMF-community relations (that the WMF still needs to work on after the monumental stupidity of the Superprotect introduction), but I also agree that proper image attribution should be more important than shiny bling, even if the shiny bling may help us get more readers/donations/anything. As one of the flagships of the free content movement, we need to lead by example, and make sure our attributions are better than what we can legally get away with, and in particular we must be better than Facebook or Google in this respect. Kusma (talk) 10:38, 21 January 2016 (UTC)Reply[reply]
This is actually an issue with the API provided by a component called Extension:PageImages that many parts of MediaWiki use including Gather. We're tracking this and aiming to fix soon. Jdlrobson (talk) 18:47, 21 January 2016 (UTC)Reply[reply]
@Jdlrobson: That should fix the non-free images issue (I hope). If this isn't deployed today, shouldn't Gather be disabled on the English Wikipedia until this is fixed? Also, do you have any thoughts on the attribution question? Kusma (talk) 21:35, 21 January 2016 (UTC)Reply[reply]
I will defer that to the WMF legal team/Reading project team as I'm just a mere developer who does what my managers ask me too. I'm just keen to get the underlying software issue fixed right now  :-) I should say this problem is not unique to Gather it's a problem with Extension:PageImages. Any application using the PageImages API (which is quite widely used in our stack) will exhibit the same problems as Gather. Thanks for bringing the issue to our attention! Jdlrobson (talk) 22:16, 21 January 2016 (UTC)Reply[reply]

RfC: Disable Gather on enwiki[edit]

There has just started an RfC on enWP about the disabling of Gather on enWP.
The RfC is not intended to be a final judgment of Gather: it only asks that Gather is totally disabled on enwiki now, and remains so until another RfC has shown community consensus that a new trial is acceptable. --Grüße vom Sänger ♫(Reden) 16:00, 26 January 2016 (UTC)Reply[reply]