Flow placed on ice

News and notes

Flow placed on ice

Flow no longer in active development

The Wikimedia Foundation collaboration team announced this week, both on Wikipedia and on the WikiTech-l mailing list, that Flow will no longer be under active development.

This news will come as a bombshell for the experienced Wikimedians who have been watching the development cycle of this project. The chaotic and disorganized nature of talk page discussions on the English Wikipedia have long concerned Wikimedians and Foundation staff alike—in a Signpost editorial published just under two months ago I wrote:

“

[Talkpages] have been around from the very beginning, but having never really substantially improved in almost a decade and a half they are today often regarded as something of a technical black sheep. There's already been one failed initiative to replace them, LiquidThreads, and another effort, Flow, has now been underway for some time, with a small number of pages currently serving as testbeds on the English Wikipedia and elsewhere. Communication using talk-pages is conceptually easy, if often messy in execution. Yet few talkpages are widely watched, and therefore, read, and so despite efforts like feedback request service there remain only a couple of on-wiki discussion points with an audience wide enough to get a point across: the village pumps come to mind, as does Jimmy Wales' talk page.

The greatest advantage of the talk pages is the fact that, being the basic venue for inter-user communication, they are accessible to all Wikipedians. The greatest disadvantage is one of presentation: lengthy posts are quickly snowballed by other lengthy posts in response, some of which are insightful, many of which are not. The lack of a visual distinction between the original author of the post and replies thereof, the blowback of the community's antiquated discussion model, causes talk page discussions to quickly degenerate into unreadability. The first and last few replies in a comment chain are far and away the most important ones, no matter the weight of their actual content, for little reason more than that they are what is most immediately read.

”

— Resident Mario, "So you want to get your message out. Where do you turn?"

Flow has been a controversial endeavor by the Foundation to thoroughly rework talk page mechanics and formatting. Its roots lay in the earlier LiquidThreads, a technical effort by WMF developers that eventually fell flat through a combination of poor technical implementation and poor community reception. At the risk of raising the eyebrows of some of our readers, the best write-up of how LiquidThreads came and went and what its relationship to Flow has been comes from Wikipediocracy: "The dream that died: Erik Möller and the WMF’s decade-long struggle for the perfect discussion system".

According to (Fuzheado), the Flow team gave an upbeat presentation on Flow's development status (video) at Wikimania just six weeks ago (though there was "hard questioning by the audience about whether the community would accept it"). Reactions are a mix of frustration that yet another effort to fix such an entrenched problem has failed and of relief that the controversial project—in the eyes of many, one of the surviving technical white elephants of the pre-Lila Tretikov era—is now apparently finito.

Flow isn't actually officially dead, according to the careful wording of the announcement. Rather, it is now out of active development pending "changes in that long-term plan". What is sidelining it now is that "article and project talk pages are used for a number of important and complex processes that those tools aren't able to handle, making Flow unsuitable for deployment on those kinds of pages." As one user pointed out:

“

The problems they outline were pretty much exactly what everyone said to them when Flow was started—you need to be able to cut and paste slabs of wikitext (or, presumably, parsoid HTML5 or whatever VE actually copies to the clipboard) from the article to the talk page, and after VE's disastrous premature introduction this was the precise thing people worried about.

”

The rest of the announcement clarifies the situation:

“

To better address the needs of our core contributors, we're now focusing our strategy on the curation, collaboration, and admin processes that take place on a variety of pages. Many of these processes use complex workarounds—templates, categories, transclusions, and lots of instructions—that turn blank wikitext talk pages into structured workflows. There are gadgets and user scripts on the larger wikis to help with some of these workflows, but these tools aren't standardized or universally available.

As these workflows grow in complexity, they become more difficult for the next generation of editors to learn and use. This has increased the workload on the people who maintain those systems today. Complex workflows are also difficult to adapt to other languages, because a wiki with thousands of articles may not need the kind of complexity that comes with managing a wiki with millions of articles. We've talked about this kind of structured workflow support at Wikimania, in user research sessions, and on wikis. It's an important area that needs a lot of discussion, exploration, and work.

Starting in October, Flow will not be in active development, as we shift the team's focus to these other priorities. We'll be helping core contributors reduce the stress of an ever-growing workload, and helping the next generation of contributors participate in those processes. Further development on these projects will be driven by the needs expressed by wiki communities.

Flow will be maintained and supported, and communities that are excited about Flow discussions will be able to use it. There are places where the discussion features are working well, with communities that are enthusiastic about them: on user talk pages, help pages, and forum/village pump-style discussion spaces. By the end of September, we'll have an opt-in Beta feature available to communities that want it, allowing users to enable Flow on their own user talk pages.

”

— Danny Horn, Collaboration team reprioritization

In the opinions of some Wikimedians, the root problem of the Wikimedia projects isn't individual problems like talk pages or templates, but rather the technical debt of a decade and a half of disorganized organic growth; the Foundation's first round of attempts at comprehensive technical improvements fell flat not because they were poorly thought out per se, but because they failed to take into account the extraordinary complexity of the use cases to which Wikimedians have adopted wikicode. SUL finalization is now complete, but plenty of other core improvements, like interwiki transclusion (to centralize template complexity) and further development of Echo notifications (to unify notification streams), remain to complete. Such core improvements may eventually make more ambitious projects like Flow manageable.

Strategic consultation concludes as community capacity building winds up

“

The goal of the consultation was to collect input on how we should respond to future trends that will affect the Wikimedia movement, and incorporate that insight into our emerging strategy.

”

In March, the WMF kicked off strategic planning consultation with the Wikimedia community. The first strategic plan was the Foundation's Goliath growth projection project, begun in 2009 and published in 2011 (Signpost coverage here, here, and elsewhere), yet it ultimately proved flat-footed at best. The Foundation began this process of self-definition anew this year (as part of a general shift towards an increasing focus on impact and impact metrics), starting with a large-scale community consultation. As we reported at the time, the WMF is trying to make the document into "what will become a discipline of ongoing strategic inquiry, assessment, and alignment. This more agile, adaptable process will directly inform and update our priorities and goals and help us maintain a strategic direction that is consistent with the Wikimedia vision, supports the Wikimedia projects, and is sensitive to the changing global environment."

The full set of findings is available, in the form of a 119-page PDF, on Commons.

The Foundation has finished digesting the outcomes of the consultation, and chief operating officer Terence Gilbey has published a blog post highlighting the findings. Part of this month's metrics meeting was dedicated to these findings, and a full deck of slides—119 pages of them—is is available on Commons.

The consultation was organized around two questions:

“

1. What major trends would you identify in addition to mobile and the next billion users?

2. Based on the future trends that you think are important, what would thriving and healthy Wikimedia projects look like?

”

Gilbey highlights the following findings:

Mobile and app: Mobile-related comments reveal an opportunity to improve our existing mobile offerings for both editors and readers and raise awareness about our native apps. Participants (mostly anonymous users) urged us to “make an app,” when one is already available for iOS and Android devices. We also saw comments that stressed the importance of mobile editing, formatting for smaller (mobile) screen sizes, article summaries for different usage patterns, and the value of “going mobile.”[3]

Editing and collaboration: In this category, we find requests to make editing simpler, ideas for enhancing collaboration among editors, suggestions for editing tools, and proposals to build editor rating and qualification programs. This is one of the few categories in which logged-in comments, at 56%, outnumber comments from anonymous and new users. This category provides valuable insight for improvements in editor support including Wikipedia’s visual editor and future projects in the newly created Community tech team, as well as potential new editor support initiatives.

Rich content: Participants requested more rich content on Wikimedia sites, suggesting more video, audio, and images. 80% of these comments were submitted by anonymous and new users. One US-based participant commented: “is there any major website in the world with less video?”

Volunteer community: We saw a particular interest in improving “community climate” in this category, with a focus on interpersonal dynamics and culture. Participants identified a need to increase diversity (in particular, gender diversity), improve processes and workflows, and address bureaucracy-related challenges. This is another category in which logged-in comments, at 54%, outnumber comments provided by anonymous and new users.

Wikimedia Foundation feedback: This category focused on the relationship between the Wikimedia Foundation and the volunteer community and includes suggestions of how the Foundation might change its practices and priorities to align with the volunteer community. These comments are from mostly logged-in users (88%), most of them highly experienced users with an average edit count of more than 64,000 edits. Suggestions included providing better support to editors in a variety of ways and continuing to ask for feedback from core community members.

Content quality (accuracy): These comments emphasized the importance of content accuracy, trustworthiness, and reliability. Comments focused on citation quality, the use of expert editors, and even restricting editing (so that “not everyone can edit”). Most (73%) of comments in this category were from anonymous and new users, signaling an opportunity to communicate to readers about the accuracy and trustworthiness of the content within Wikipedia and sister projects.

Education and universities: These comments reflected both a concern about the perception of Wikipedia as a (non)credible source for academic inquiry, and also recognition of the growing opportunity for Wikimedia to extend its content, brand, and global presence into online education by developing courses, curricula, and partnering with other online educational resources. 76% of the comments in this category came from anonymous and new users, whereas only 24% originated from logged-in users.

Translation and languages: We saw a collective interest in this category from logged-in, anonymous, and new users. Key suggestions included a focus increasing translation capabilities and tool, expanding into more languages, and developing the ability to easily translate across projects. These comments validate the need for the Content Translation tool, which is now available on 224 language-versions of Wikipedia as a beta feature.

In related news, the Foundation is now engaging in what it calls a community capacity development project. According to an an email to the mailing list posted by the WMF's senior program officer, emerging Wikimedia communities Asaf Bartov, the Foundation is allocating staff time to "deliberate capacity-development projects with interested communities in six capacity areas: community governance; conflict management; on-wiki technical skills; new contributor engagement and growth; partnerships; [and] communications". "Community capacity" is defined as "the ability of a community to achieve ... very diverse [goals that] span issues that affect one or all Wikimedia communities." It is, in effect, a trial of a more hands-on approach on the part of the Wikimedia Foundation in recruiting ideas from the community, following along the lines of earlier breakout efforts, most prominently this year's "Inspire" campaign.

Brief notes

Philippe Beaudette, the now-outgoing director of community advocacy at the Wikimedia Foundation.

Philippe Beaudette leaving the Foundation: Long-time Wikimedian and director of community advocacy Philippe Beaudette is leaving the Foundation. Beaudette was brought on in 2009 as a facilitator for the Strategic Planning project that was ongoing at the time—by his own count, he was employee number 16—making him one of the most senior of the officers still with the WMF. In the intervening six years, Beaudette served variously as facilitator of strategic planning, head of fundraising (as part of the 2010 Annual Giving campaign), and head of Legal And Community Advocacy. Beaudette has been an administrator on the English Wikipedia since 2007.
IEG round: The Individual Engagement Grants grantmaking program is holding a submission round from August 31 to September 29. The IEG program is the funding platform of choice for small individual or team-based research, development, and engagement projects. Two of the program's grant officers will he holding Google Hangouts—the first on September 8, to answer proposal questions and to help polish grants that could benefit from feedback and development at the IdeaLab.
Metrics: The next WMF Metrics and activities meeting will be on Thursday, September 3.
Develop the Digital Library Card Platform: The Wikimedia Library program has been growing at a fast clip and now finds itself in need of a stronger core platform. Last week saw the announcement of the opening of a position for a developer experienced with web applications and open-development frameworks who can build a Wikimedia Library "Digital Library Card Platform tool". A lengthier description as well as a point of contact for interested parties is available here.
Mobile legalese: The WMF has published a set of guidelines for mobile developers using the Wikimedia trademark, presented in condensed form in a recent blog post.
August in Edu: The August edition of the Education Newsletter has now been published.
RfC on the 2015 ArbCom election: A request for comment on the upcoming Arbitration Committee election has been opened. The RfC was created to "provide an opportunity to amend the structure, rules, and procedures of the December 2015 English Wikipedia Arbitration Committee election and resolve any issues not covered by existing rules."
WLM banner saga continues: A referral for comment has been spun up on the meta-wiki regarding the time conflict between Foundation fundraising and the Wiki Loves Monuments movement. Given the lopsided nature of the vote, as well as the loaded questions being asked by the RfC organizers, your author considers it yet more evidence of the inflammatory feedback loop between Foundation non-communication and community worked-up-ness. For more coverage see this week's Discussion report.
New administrators: Wbm1058 is this week's newest administrator.
100K: The Belarusian Wikipedia reached 100,000 articles this week. Keeping in the tradition of things the encyclopedia uploaded a unique 100,000-themed take on the Wikipedia logo.

← Previous "News and notes"

Next "News and notes" →

In this issue

2 September 2015 (all comments)

Special report

News and notes

Discussion report

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

So, now it's time to allow VisualEditor in Talk. Not that VE is ready for prime time, but it is developing. Jim.henderson (talk) 23:37, 3 September 2015 (UTC)[reply]
Flow is the Duke Nukem Forever of MediaWiki extensions. It was developed with no clear plan or requirements, without any regard whatsoever to the purpose of Wikipedia or whether it was actually wanted and prematurely deployed in an unacceptably buggy and incomplete state. It is symbolic of everything that is wrong with WMF "Features" technology management. Good riddance. MER-C 00:59, 4 September 2015 (UTC)[reply]

The effective termination of Flow is probably the best news about Wikipedia that I think we're gonna hear this year. Anybody who witnessed the destruction wrought upon the "Off-Wiki" website through the introduction of LiquidThreads (forerunner of Flow) can not have been anything but terrified about the forced adoption of this software. It was an asteroid on collision course with English-Wikipedia — now fortunately diverted. Kudos to Lila T. and the new cooperative attitude emanating from San Francisco. Carrite (talk) 01:05, 4 September 2015 (UTC)[reply]
As mentioned in the article there is a lot of great stuff we can do to improve our current talk page system that with effort could be easily rolled out in a timely manner. Doc James (talk · contribs · email) 01:22, 4 September 2015 (UTC)[reply]
I am not sad to see Flow finally being practically being the given the coup de grâce. I am sad about all the time and money that has been wasted on it while other far more serious priorities have been more quietly swept under the carpet. --Kudpung กุดผึ้ง (talk) 03:29, 4 September 2015 (UTC)[reply]

Thank you for mentioning my departure. That's very kind. A couple of little corrections: I wasn't head of fundraising, I was head of the annual giving campaign (Zack Exley was the closest thing to Head of Fundraising, which included other teams as well as mine - Foundation relations, Major Gifts, etc.). I was also not Head of Legal and Community Advocacy - I was Director of Community Advocacy, with no authority over the "L" in "LCA" at all. Nothing like titles and job functions that are as clear as mud, huh? Thanks again. Philippe Beaudette, Wikimedia Foundation (talk) 03:43, 4 September 2015 (UTC)[reply]
- User:Philippe (WMF) we will miss you from your foundation roll but hope to see you just as engaged with the community as every, just now as a fellow community member. Doc James (talk · contribs · email) 03:57, 4 September 2015 (UTC)[reply]

but plenty of other core improvements, like interwiki transclusion (to centralize template complexity) and further development of Echo notifications (to unify notification streams), remain to complete <-- They remain to complete, because essentially nobody is working on them. They certainly aren't the sort of thing you can do in a weekend, but its not like they are impossible tasks either. In fact, I would consider them significantly less ambitious then a lot of things that the foundation does. If they were considered high priority things, with a team actively working on them, both of those could probably be done in a couple months. I'm also not sure how they would really help flow's nebulous goal (or past goal, I don't think its been a "real" goal of flow for quite some time now) of being a replacement for everything involving a discussion. Bawolff (talk) 07:29, 4 September 2015 (UTC)[reply]
Imma offer a somewhat contrarian position here. I've always been concerned about how the current Talk system is trying to use a free-form wiki page for mostly structured discussions. Most of the world does not use free form pages for structured discussions, which may say something about its suitability. Easy to put stuff in the wrong place of the conversation, the need for signing, easy to get confused about where X's post stops and Y's post starts and so on. Putting up header messages is really the main shortcoming in Flow, but a remediable one. Interwiki transclusion and interwiki notifications would be a great thing to work on, now. Also, farewell to Philippe. Jo-Jo Eumerus (talk, contributions) 09:47, 4 September 2015 (UTC)[reply]
Hooray! Now I don't need to invent creative workarounds for Flow! A shame about all the time and money spent on it, though... Double sharp (talk) 13:35, 4 September 2015 (UTC)[reply]
When there is an announcement to make about a project which involves test subjects, volunteers who participated in the research should be the first to be notified. I am a participant in a WikiProject where the members gave consent to test Flow so that the developers could get user data. Using Flow caused problems which volunteers would not have had otherwise, but I am glad that the developers got some test data. I have no opinion about what is reported in this article - Flow was a nice idea, and am I am sure the decision to stop development was thoughtful.

I feel that the Flow developers should have notified the communities using Flow first, so that those communities could be prepared to have discussions about their future relationship with Flow when more public announcements are made and the wider community is ready to discuss these things. This is true of research generally - volunteer research participants should get priority notification of news about research in which they participate and which affects them. As things are, I have doubts that forums that used Flow would agree to continue using Flow if the developing team does not want userdata any longer. I posted a message to the Flow developer page, but I might have preferred that developers come to the forums where they asked to test Flow and give an option to revert the format to the usual style and let the community decide what to do next.

I am going to signal this at meta:Research:Committee also. When research happens and involves human participants there should be awareness built into the research of how much volunteer time is being consumed and how the impact to volunteers can be minimized. Blue Rasberry (talk) 13:52, 4 September 2015 (UTC)[reply]

Re "We're now focusing our strategy on the curation, collaboration, and admin processes that take place on a variety of pages", WP:DRN and WP:FIX are two such pages that could really use some help from the WMF. How do I inform the appropriate people at the WMF that such a need exists so someone can prioritize it among the other requests for help? --Guy Macon (talk) 00:46, 5 September 2015 (UTC)[reply]

I for one am gutted at the news that development is to stop on Flow. I think it's an excellent tool that works much better than talk pages, and in the vast majority of cases is already at a point where it would be beneficial to replace them with Flow. I'm really pleased to hear there will still be an opt-in to convert user talk pages to Flow, I'll certainly be signing up for that. Waggers TALK 11:59, 7 September 2015 (UTC)[reply]
About
There's already been one failed initiative to replace them, LiquidThreads...
You should actually count that as two initiatives, since mw:LiquidThreads 3.0 also didn't succeed. Helder 16:44, 9 September 2015 (UTC)[reply]

While I agree that the talk page system could be made more user-friendly for newbies, I am glad Flow is on the chopping block. Modeling everything on WP to be more like Facebook is not the way to go.

Although we've already had a final heckling of Philippe on the functionaries mailing list, I'd like to publicly thank him for all his work over he years. He did a lot for this project, answering emergency emails at 3 AM, helping us deal with the worst of the worst of abusive users, and keeping the functionaries informed about things that affected us. Although I have no doubt His replacement is up to the task, he will be missed. Beeblebrox (talk) 22:14, 10 September 2015 (UTC)[reply]

@Resident Mario: Thanks for linking to my Wikipediocracy piece about LQT and Flow. — Scott • talk 09:19, 11 September 2015 (UTC)[reply]

I am shocked that we are now permitting linkage to a hate & attack site! - 2001:558:1400:10:502C:71A6:6A0:2CFD (talk) 13:10, 11 September 2015 (UTC)[reply]

Not everything Wikipediocracy ever published is bad. ResMar 13:15, 11 September 2015 (UTC)[reply]

I think that Mr 2001, the Comcast ~~employee~~, sorry user from Philadelphia might not need to be told that.--92.238.57.40 (talk) 13:27, 11 September 2015 (UTC)[reply]

What is this place coming to? 122.111.234.200 (talk) 20:06, 11 September 2015 (UTC)[reply]

This is more WMF "Communication problems". There is a temporary pause on Flow "discussion features", but the WMF is still developing Flow and still planning to deploy. In fact the WMF Executive Director announced an intent to deploy it to replace her own Talk page, and still plans on "eventually getting it ready for prime-time". The announcement that they were shifting to work on stuff to help the community was also very misleading. I spoke to the project manager. What they meant is that they are working on a FLOW-ONLY project. (Called workflow.) He said that many of the large projects will get no benefit from it..... because we haven't converted our pages to Flow. When this was initially announced he promised the project would be driven by the needs expressed by the communities. I said obviously the new workflow project should be compatible with existing pages. No dice. I asked if he would respect an RFC saying we need workflow to be compatible with existing pages. I was told that an RFC wasn't needed, the community were a bunch of change-averse luddites, and basically that those Community Needs would be ignored. I then asked if I brought him multiple RFCs (or a multi-wiki RFC) resonably representing the broader Community, would he respect that. So far no response, and it doesn't look promising. The WMF wants our existing pages GONE, still intents to deploy Flow, and is developing new projects as Flow-only. If we don't want Flow, we get a 'screw-you'. The AGF here is that the WMF thinks it's doing the right thing..... maybe their announcements were unintentionally misleading.... but nonetheless everyone (including the author of this Signpost story) were grossly misled by the WMF announcments. Alsee (talk) 23:38, 17 September 2015 (UTC)[reply]

P.S. I tested Flow a bit. Every time I touch it I run into critical problems. Just to cite SOME of the issues I've reported, copy-paste mangles content. Trying to UNDO an edit can mangle content. Flow has two edit modes (Visual and Wikitext)... merely switching between the two modes can mangle content. The Flow discussion threading model mixes top-posting with bottom-posting, which turns larger discussions into incomprehensible spaghetti. Flow expands discussions to more than twice the vertical size (the Flow FAQ has a section "Why does it look like Facebook", explaining this is because Facebook-type usability testing says it's better). Initially Flow developers were simply told to build a chatboard - and they succeed at that. But now Flow is a pile of awful kludges trying to upgrade that chatboard into something editors can use. The kludges are bursting at the seams. It's a disaster. It seems Flow actually worked BETTER when I tested it a year ago. Alsee (talk) 23:53, 17 September 2015 (UTC)[reply]

Want the latest Signpost delivered to your talk page each month?

Home

About