On 10/31/2011 11:04 AM, David Levy wrote:

Erik Moeller wrote:

...

As a matter of general practice, the Wikimedia Foundation aims to be responsive to the community both before and after the deployment of software, but it doesn't obtain community consensus before deploying software which it would like to deploy on its sites and services, nor does it necessarily write or deploy software changes if a consensus to do so exists.

Brandon Harris explicitly stated that "the policy for deployment of the tool is that it is by request only, and the requesting wiki must ... show community consensus."

http://lists.wikimedia.org/pipermail/foundation-l/2011-October/070062.html

That leaves three possibilities:

A) Community consensus was demonstrated at the English Wikipedia. B) The WikiLove deployment policy was violated. C) The above statement by Brandon Harris is incorrect.

If I understand correctly, the English Wikipedia is the main test deployment for this as an experimental feature. While the feature remains experimental, additional deployments to other wikis would only happen if requested by community consensus. At some level, it would not make sense to insist that consensus is required prior to conducting any experiment, as that effectively defeats the ability to experiment.

--Michael Snow

On Mon, Oct 31, 2011 at 19:36, Erik Moeller erik@wikimedia.org wrote:

On Mon, Oct 31, 2011 at 11:22 AM, Michael Snow wikipedia@frontier.com wrote:

...

If I understand correctly, the English Wikipedia is the main test deployment for this as an experimental feature. While the feature remains experimental, additional deployments to other wikis would only happen if requested by community consensus.

That's right. Because we're not actively organizing or vetting any efforts to localize the feature beyond its initial test deployment, we can't deploy to other languages unless there's a clear, proposed configuration and a consensus to use it. Given the still experimental nature of the feature, and the relatively high cost to manage a community-wide change, that's purely a pragmatic choice. We've made the same choice for ArticleFeedback and other experimental features, and will likely do so with others.

There are two issues here:

First, it's not the same to implement (a) something really needed (let's say, optional WYSIWYG editor when it would be done, hopefully before the next decade), (b) something about editors don't care, (c) something which has potential to irritate editors because of various reasons. (Wikilove irritates me, for example, because it was promoted as "improvement", "new cool thing", while it's basically useless, irritates others because of other things and talks a lot about disconnect between WMF staff and reality.)

In the case (a) I don't think that any sane person would say anything against implementation. Of course that you should implement WYSIWYG editor ASAP. The case (b) is not an issue because nobody cares about it.

The case (c) could be easily predicted (if you WMF staff is not able to predict such things, you should consider employing an expert in predicting things; I could suggest you a couple of good shamans). And it's your job to prevent conflicts, not to make them because a couple of you think that something "is really cool" and that "it could pass" and "if not, we'll think later". There are very straight-forward options for that, including the one below.

Second issue is related to the position of English Wikipedia. Yes, it's obvious that you need to show some feature somewhere and that English Wikipedia is the best place to do that. However, instead of just putting it and making excuses of various kinds, you should make clear that it's used as an example for limited period of time (ten years, if you want), after which community would be able to decide does it want it or not. Otherwise, it doesn't just sound arrogant, but it *is* arrogant behavior.

On 10/31/11 11:08 AM, Nathan wrote:

It sounds like Brandon misunderstood the criteria for deploying WikiLove on various projects. Can you give us the correct description of how that is being handled?

Well. I wrote the criteria for additional deployment.

You can read them here:

http://www.mediawiki.org/wiki/WikiLove#Requests_for_Additional_Deployment

On Mon, Oct 31, 2011 at 11:08 AM, Nathan nawrich@gmail.com wrote:

That's a pretty bold statement for the WMF to make - "If you don't trust the WMF, don't contribute to WMF projects." Are you sure that's what you meant?

Hi Nathan,

let me try to clarify what I mean by trust in this context. We can, indeed must, talk very openly about what works and what doesn't, and whether we're doing the right kinds of things. That discourse, and the readiness to engage in it and to change course for the right reasons, is key for a relationship based on mutual trust to work.

It's easy for us to accidentally send mixed messages, though, as this thread has shown. Because so many things are done in response to community consensus, there may be an expectation that this is always the case, and that that's just how we work. Change in Wikimedia projects has, however, always been a continuous process of give and take, with a certain element of arbitrariness, seeking to find the right balance between acceptance and progress, and "being bold" to try new things. That process can be very messy, as the image filter discussion has shown.

So, what I'm saying is that if you (generic "you") believe, for whichever reason (lack of trust, philosophical reasons, or whatever), that WMF shouldn't be permitted to make meaningful changes in Wikimedia projects without obtaining upfront community consensus, you'll probably find more satisfaction and joy in volunteering in a different context. That's not how WMF projects operate, and they are very unlikely to ever do so, for reasons that have been articulated here and elsewhere. Indeed, it's part of WMF's understanding of itself that part of its job is to continually challenge existing agreements and practices in order to support positive change.

Erik