I of course cannot speak for the Foundation. I only write this in the view of a volunteer dev, like many others. That statement was written a long time ago when Mediawiki was simply the software that runs Wikipedia. It's now 2009, and Mediawiki is still the software that runs Wikipedia. That being said, our outside user base has grown massively in this time. A good number of our bug reports and patches come from outside users, not wikis within the WMF. That's all fine and dandy, but our number one goal is still (admittedly or not) to keep developing for Wikipedia. I of course support full consultation with the wikis when it is beneficial to do so. Simple bugfixes or enhancements don't need massive pre-announcement and input. It slows down the development lifecycle for everyone. Most devs don't want to be involved in massive enwiki debates over where to put a link: we just want your final consensus on what you want done (and that itself can be very time consuming). Larger impact things (like the retooling of wikitext) definitely need wider input than just wikitech-l. I believe that the WMF community and wider wiki community should be solicited for such wide-sweeping changes. Tangentally, I think we all as a wiki community need to standardize "What is wikitext" in a formal way, but that's another discussion. At this day and age, I would hope silly feature hacks for things only wanted by one wiki could be avoided. We've had quite a bit of feature-cruft over the years, and a lot of these things probably would've been better as extensions to begin with. In short: I as a developer welcome all input from the wiki community (both WMF and not), and I highly encourage those who share an interest in the direction of the software (not everyone does) to get involved. I'm not going to track you down and poll everyone around you, but I will certainly listen carefully to your ideas. Always, Chad On Jul 1, 2009 1:16 AM, "Brian" <Brian.Mingus(a)colorado.edu> wrote: Going forward, how does the Foundation plan to make large changes to the software in full consultation with the community consensus? Is the assumption that all of the members of the community who are knowledgeable and interested have already signed up to the relevant mailing lists and all that is needed is to send out a quick 'ping' and get their thoughts? What constitutes the community when it comes to the software? Or is this just a guideline that has been on Jimbo's user page for many years which is not really relevant since laymen should not really be involved in technical decisions? Is there anyone at the Foundation who currently takes this principle seriously? Honestly? What about the developers - are they aware of and actively engaged in implementing this principle? Does the Foundation feel that it doesn't actually need to consult the community? It can determine the technically best solution for the projects and then implement it without soliciting feedback from as many people as possible? What would constitute due diligence in contacting the community? For example, suppose that the Foundation had determined that there were 5 really good solutions to a problem in the software and that they take full consultation seriously. Could you then present those 5 solutions to the community en masse using a survey, analyze the results and choose a winner (or have a runoff?). How large of a change to the software requires full consultation? After consulting the community, does the Foundation feel it is within its power to then choose something different? Does the Foundation take the requirement that all changes to the software must be gradual and reversible seriously, or not? What does that mean to you? Thanks, Brian _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l

On Wed, Jul 1, 2009 at 1:16 AM, Brian wrote: > Is the assumption that all of the members of the community who are > knowledgeable and interested have already signed up to the relevant mailing > lists and all that is needed is to send out a quick 'ping' and get their > thoughts? Yes, IMO (as a volunteer dev). If something is expected to be controversial on non-technical grounds, there's normally a per-community decision, like with FlaggedRevs or whatnot. The overwhelming majority of technical work comprises straightforward enhancements and bug fixes that only really deserves technical attention. Users who are interested can sign up to wikitech-l and hang out in #mediawiki. Those who aren't can just use the software. > Or is this just a guideline that has been on Jimbo's user page for many > years which is not really relevant Yes. > How large of a change to the software requires full consultation? It's not about large, it's about the effect it has on users. There are enormous overhauls like the new video upload system that don't need to be discussed with the community at all, because everyone agrees that they're wanted. On the other hand, there are plenty of one-line changes that would require community consensus (like, say, giving all users rollback by default). > After consulting the community, does the Foundation feel it is within its > power to then choose something different? I can't speak for Wikimedia, but I don't see how it could possibly be considered outside the Foundation's power to ignore the community. It owns the site. It can and does overrule individual communities sometimes, in technical matters and non-technical alike. For instance, it imposes its copyright policies regardless of community consensus (and some communities don't like those policies at all). An upcoming technical change that will probably be very controversial is deployment of the Vector skin by default -- I predict a lot of people will complain about lack of consensus and be very politely told "too bad, we know better because we just spent a million dollars on a usability study". > Does the Foundation take the requirement that all changes to the software > must be gradual and reversible seriously, or not? What does that mean to > you? It doesn't mean anything to me. Many large software changes need to be deployed all at once, and many aren't easily reversible. That's life. []

On Wed, Jul 1, 2009 at 8:32 PM, Jimmy Wales <jwales(a)wikia-inc.com> wrote: One of my favorite suggestions, from Erik, is that we use IdeaTorrent ( http://brainstorm.ubuntu.com/ ) in order to provide a single place for users to engage in very important discussions about all manner of issues relating to the community. Right now there is simply no way for our widely disparate community of users, who have expertise in every area imaginable and whose collective input is extremely valuable, to come together and have a conversation. It takes someone such as yourself to champion an idea to the community and present it in its best form and make sure that the best of the arguments from both sides are heard from users. As the projects have grown and as they have become more centrally managed in a top down fashion it has become increasingly difficult for ideas to percolate from the bottom up. How can a user with a great idea on one wiki present it and be sure that users from the other wikis and the WMF see it? Likewise, how can the Foundation ask questions of the *entire* community? Neither users or the Foundation have a voice that can reach everyone (fundraising and the like are an exception). There isn't a plausible conduit through which we can present and receive ideas and those ideas are considered on an equal basis with all other ideas and then refined and improved by the will of the community and ultimately implemented (by a volunteer or the WMF). Regarding the software, I think it's great to hold a conversation on wikitech-l about the best way to replace ParserFunctions. Of course, the way we got ParserFunctions was through a conversation on wikitech-l which entertained a few ideas but ultimately did not have the wider goals of the community in mind due to the narrow scope of the discussions. Usability and encyclopedia writing were not concerns, CPU cycles was. The justification was, and continues to be, well, there is obviously a problem here. Therefore we, the code writers, have free license to develop a new solution, ask our friends in IRC if it looks nice, and then put it on the live sites. It's not even clear how you could extract a consensus from wikitech-l if it were there. If you take fully consulting the community consensus seriously then there is a very different design model that then leads to development. In this method we have a plausible way of asking a large number of *editors and users* what is wrong with the software. You have to get many of the people who actually edit the encyclopedia a lot and have something to say about what's wrong with it and what's right with it in the same place fully engaged with each other. Right now we do not have tools that facilitate this. Article talk pages are simply not it. Meta is not it - the people aren't there. The mailing lists aren't it - the people aren't here. We represent a tiny minority of the community and a minority of the total number of people who would, if they were afforded the opportunity, have an opinion worth hearing. If you look at the number of people engaged in any conversation which will have a serious impact on all of the projects and then compare that number to any measure of active editors and contributors you will see that it is shockingly small. I encourage the WMF to make that ratio as large as possible, and I suggest that the larger you make it the more we will all benefit. I get the feeling that many people look at full consultation as a lot of really hard work. I think that's wrong - we are supposed to be leveraging the power of communities. The WMF has the power to enable a community to come together and form a consensus by bringing their attention all to the same place. I think that until something like that happens full consultation is more of a dream that many people aren't even trying to realize and changes will continue to be made to the software and otherwise which aren't really in the right direction. For example, it's not clear to everyone that Wikipedia even needs a programming language. I don't know if it does or not. There are a lot of things to take into consideration, such as usability, readability of the main article namespace, duplication of content, ease of more sophisticated editing, and issues you or I might not even think of! Adding a programming language is not a magic bullet to these issues. It could in fact be that templates and the way we work with article content in the first place needs to be entirely rethought. This is not a conversation that should be limited to wikitech-l. In fact, editors might have a more useful opinion. But in the current system their opinions won't be sought out as the decision to do it was entirely top down. Lastly, I do not consider a wide distaste for the look of ParserFunctions to be a sanction for a new programming language. ParserFunctions was added because a few users decided to abuse templates in order to get if-like functionality, but it was not done in consultation with the community. Thus, at the very least, I suggest we go back to the community and ask them, from the perspective of just having vanilla templates, how we can improve the software to make Wikipedia and the other projects better. And I suggest that we solicit feedback from as many people as we possibly can - fully consult the community consensus!