(English) Wikinews suffers from three main problems that are related to the software we use.
1) MediaWiki is not a good CMS, in the traditional sense -- it does not easily support workflows, and depends on humans to adhere to a well-defined process. 2) MediaWiki does not support RSS feeds. Today's news junkies depend on RSS to read their news: our ability offer them these feeds is near zero. 3) MediaWiki does not support dynamic listing. If I tag a news story as Category:Canada and Category:Elections, there is no way currently for the system to automatically publish this list to a dated list of stories about Canada, or Elections, or Elections in Canada. In other words, MediaWiki doesn't allow for server-side views on data that are unions of categories.
The combination of these three factors to varying degrees has a profound effect on our ability to scale.
-Not having a CMS, we have to teach every user how to follow our process -- losing quite a bit of time on the effort. - Not having configurable, high-quality RSS feeds means that bloggers and news pundits won't readily access our content. - Not having dynamic lists means that for every story that we write, we have to edit at _least_ five other pages -- or as many as 20. This leads us to lose a significant amount of time and effort, or delivers inconsistent results if not everyone follows the process, or delivers an incomplete news site if we follow a simpler process.
Wikinews and our current active contributors have shown that we can put together about 20 stories a day on a somewhat sustained basis. I think that we are very near our limit, however, and that further growth is unsustainable, due to the limitations of the software. Unlike Wikipedia where you can edit a story and have someone link to it tomorrow, or three weeks from now, on Wikinews all the links have to be in place at the right time -- or face irrelevance.
There are other need-to-have features, of course -- but these are the ones that will likely determine the survival and growth of the project in the long-term, and ones that are serious enough changes from the current state of MediaWiki to warrant this discussion.
We have made strides in resolving this. Bugzilla bug# 1411 has had an extension implemented that will partially solve our Dynamic List problem, but was shot down by developers as unacceptable (though I suspect it's quite acceptable to a smaller site like Wikinews, even if unacceptable to a large site like en.wikipedia). We have tried putting together external RSS feeds, but with only limited success.
I am ready to make a more detailed proposal regarding the specific work that needs to get done, but I am glad that Gerard started the discussion in the meanwhile. Wikinews (and probably other projects) needs certain software changes to grow. In the current environment, there is a catch-22 -- we have to grow big enough to be important to developers to have big changes made to the software, but we need these changes made to grow. Not all of these changes are even useful in Wikipedia. Not all of these changes _should_ or _could_ even be used in Wikipedia.
Amgine brought up hardware or software forking. The discussions (held mainly informally, on IRC) focused on perhaps donating hardware to the Foundation to run Wikinews on separate servers at the current datacenter: allowing us to experiment in ways that the current development and release process might not be happy with. The software discussion, likewise, faced the fact that we either have to have someone on our team become a core developer, or we have to find alternate software on which to run Wikinews -- or get enough support from the Foundation to make the changes we need.
Sorry to be so verbose, but I want to chime in to explain our state of things more fully.
-ilya (n:User:IlyaHaykinson)
On Mon, 28 Mar 2005 17:55:37 +0200, Anthere anthere9@yahoo.com wrote:
Amgine a écrit:
In wikinews there has been an ongoing discussion regarding the unresponsiveness of the mediawiki team to issues considered critical to the project. This has now become a discussion regarding a fork of both hardware and software, and contributors are actively working toward doing so. I don't think this is the best choice, but it is beginning to look like the only choice.
Amgine
I do not think finding hardware to host a project this size would be very problematic in the short run, but who would take care of it ? And mostly, if you think the current team is unresponsive to current wikinews software needs, who do you think will take care of modifying the MediaWiki software outside of the current team ? Is it perceived it would be more efficient to have a software fork or even to use another software ?
Incidently, though I often heard of some software features for the wiktionary and species project, which features are necessary to wikinews and where has this been discussed ?
Ant
foundation-l mailing list foundation-l@wikimedia.org http://mail.wikipedia.org/mailman/listinfo/foundation-l