Incubator talk:Policy/First draft
I disagree with "How does a New Project Get Created on Wikimedia Incubator?". I think a new language should be started here before request on Meta; the Meta request should serve as confirmation that a test-Wikipedia is active and should get its own subdomain. The situation is a little different for new project, but I guess we could apply the same way, as any project with a working demo could be considered better than without it - and the demo should be on Incubator. — Timichal • 08:35, 26. Jul 2006
- But there ought to be some kind of voting or decision before something gets started here, or we'll be overrun with rubbish ideas which only have 2 pages. Maybe make sure that at least a certain number of people have agreed to help before allowing it to be set up. We could even have a proposal page here where people could vote. Dbmag9 11:36, 27 July 2006 (UTC)
Trimming the Fat of Spurious Proposals and other thoughts
When I set up the original Wikikernel new project policy, it was done both to satisfy some issues raised by User:Angela, as well as from experience I have had trying to get a couple of different project up to sister project status, notably Wikiversity and Wikistandards. Both of these have gone through a formal voting process, and some signficant ink was spilled for both of them regarding both the voting process and some very strong feelings both for and against each project and some very legitimate issues that were brought up about their creation.
I have been a long-time advocate of new project development reform, and this policy page lies right at the crossroads to effectively performing that task. On average about two or three serious new project proposals are create each month, with about five or six each year that have had participants go so far as to either create a new Wikicity (now Wikia) or even buy their own web hosting services and start the project as a stand-alone group willing to go it alone on the remote hopes that eventually they will merge back into the Wikimedia project fold.
The only successful project that went cold turkey and started as an independent project but has "come back" to become an official Wikimedia project is WiktionaryZ, which BTW was something that has had long-time support anyway and was more technically inclined and not so much content development. Even then, its official status is somewhat dubious and still remains as an experimental server.
Of those that have tried to duke it out and remain within the Wikimedia family of projects, there are two significant projects that have semi-successfully gone through the whole new project proposal and become full sister projects: Wikinews and Wikiversity. The process that all other Wikimedia projects went through, including this incubator project (unfortunately) was merely the whim and thought of either a board member or somebody on the development team that merely decided it was time and created the projects. Wikibooks did start out as a project proposal after a fashion, but the WMF has changed considerably in the past three years... in fact it didn't even exist back then. Wikisource, Wiktionary, and Wikiquote have all been sort of haphazard in their creation as well, and the worst example was the creation of Wikispecies, which has been derided numerous times on Foundation-l.
Wikispecies has been derided so much that it is used as the benchmark for what a new project proposal should never be like. Board members wanted it to be made clear that Wikispecies should not be used as justification for creating a new Wikimedia project, and that any attempt to create a new project should have a much more defined scope and user support base before it is ever created. Other regular participants to Foundation-l and other forums have expressed similar thoughts.
I'm glad that the Incubator Wiki is now available to help bridge this gap. Please note meta:New project proposals group and previous inititives to get this working, and in particular note Angela's comment she made when she removed her name from the list of participants on that page (in the history of the page): "Too many of these should never be projects and having this group just gives people the idea new project proposals are taken seriously, when I don't believe that is currently the case."
I think it speaks volumes toward board attitude toward new project proposals. We need to have a culling procedure that removes silly ideas, but the current process is so bogged down and harsh that I have been openly critical of the Wikimedia Foundation Board of Trustees on this point and have suggested that WMF policy should be: No new Wikimedia project will ever be accepted. Thank you for trying, but instead try to fit within the framework of existing Wikimedia projects or try your idea out on Wikicities. Thank you for your time.
It isn't that harsh as some work has gone into developing some other new projects, but there are still some problems. Certainly we need to find some compromise between that ultimately unforgiving sentiment of no new Wikimedia project ideas at all and permitting every silly idea webspace as a free webhosting service. I suggest the bar has been for far too long pushed toward killing every new idea, but I don't want it to swing the other way immediately and suggest that every idea should be accepted either. --Robert Horning 22:51, 11 August 2006 (UTC)
- Interesting thoughts. I agree that the current process of setting up a new project is undefined and troublesome. Currently users have to make a proposal page, list it on Proposals for new projects, read the policy page, send messages to foundation-l, try and raise support, come up with policies, well-defined rules, users who will be contributing... and after all this hope that the board approves the project.
- Incubator has tremendous scope for eradicating (some of) these problems. If we have a lowish bar for new projects here (10 users who will contribute, for example), then demos can be set up in tandem with a proposal page and notification at Meta. Discussion will have something concrete to work on, and after a certain period users can start to think of moving the test into a full domain.
- The new project policy needs improvement. Incubator can help with this improvement. In the meantime, feel free to change the Incubator policy to show what needs to be done.
- Thanks! Dbmag9 09:13, 12 August 2006 (UTC)
The fact that this is proposed policy is very important. It should be included in mediawiki:sitenotice.
I have a number of objections, questions, etc., and I feel most of these should be discussed and, if no consensus can be reached, put to a vote (eventually -- after discussion is tried, not right away).
- Why can't people create their own tests? Why do people need to make requests? What's wrong with the old system that was carried over from meta, wherein people can start a test whenever they want? I understand requiring requests for new _project_ tests, but not for new language tests for pre-existing projects.
- Because, to be useful, Incubator needs to be organised. A test which is in a language with one speaker and which takes up the language code for a genuine language does not benefit anyone. If users must go through any kind of process to create the test, the resulting tests will be overall of a lot more use. It forces users to think seriously about whether the test is suitable, and also prevents situations in which a test is created for a language which already exists.
- Well, this can easily be solved by creating a rule requiring people to use the proper code for their language, and if their language doesn't have a code, establishes a process for selecting a code. The Incubator concept has worked just fine since its original conception without the requirement of test-approval. I don't see why it needs to start now. Also, if a test is created for a language which already exists, this can easily be remedied.
- The requirements for establishment of a new test are far too stringent. The vast majority of tests so far (including successful ones) were started by one or two individuals, not 5 committed users.
- The requirements are up for debate, but are at present being considered as part of the process for making a new language, which is being decided by the New Languages Subcommittee of the SPC. The figure of 5 is probationary, and will indubitably be subject to change.
- I can understand removal of inactive tests from the front page, but what is the motive for total deletion? What purpose does it serve?
- It serves the purpose of maintaining the Incubator's organisation and usefulness. A test which has been untouched for several months simply clutters up the Incubator, and makes the hierarchical organisation harder to navigate.
- Really, though, what is the purpose of the hierarchial organisation? If someone wants to access an active test, they can use the main page. And it would be perfectly acceptable to, say, remove the categories from the pages of inactive tests.
- There should be deletion guidelines for tests. For example, if somebody nominated en:Internet for deletion, it would lose because it meets 0 of the generally accepted criteria. Some starting points might be:
- Does it have more than one user?
- Is it a "real" language whose existence is documented outside of Incubator? (note that this does not exclude conlangs, they just have to be documented outside of Incubator).
- Does it house primarily or exclusively content that could go equally well at another Wikimedia project?
- It makes sense, based on these points, that it should fail two or more of the questions before deletion is a realistic option. If it isn't documented beyond Incubator, but it has 7 users, I don't think it makes sense to delete it, although it would certainly be a problem and an investigation would be warranted.
- There is already a proposed process for nominating tests for deletion. Most of the criteria are already there.
--Node ue 02:44, 8 October 2006 (UTC)
- I've answered your points above in turn, in red. Note that most of the time I have simply contradicted what you say; this is deliberate. Overall the policy is unlikely to be changed hugely, not because it is some kind of fatwa but because it is in line with the will of the powers that be, however almost all of the actual figures are up for debate. Thank you. Dbmag9 13:41, 8 October 2006 (UTC)
On IRC there was mentioned that Incubator should be used for further internationalization of MediaWiki, and support the development of new
MessageXxx.php files. At first sight that seems to be going together well with the development of new language project for existing WikiMedia Foundation projects. Open questions:
- Where to put these i18n support pages? Obviously, the're not part of an existing test.
- To what extent should support go? E.g. can we install special MediaWiki software extensions facilitating language interface development?
- How shall effords here relate to existing work on betawiki?
--Purodha 12:07, 19 February 2007 (UTC)