Rowan Collins schrieb:
On Wed, 12 Jan 2005 20:51:32 +0100, Frando frando@xcite-online.de wrote:
My idea is to transform the discussion pages to a real forum, as known from phpBB, Woltlab Burning Board or vBulletin.
I think everyone will agree that some move to incorporate the advantages of forum systems within the context of the wiki would be a good idea. Let's face it, wikis just *aren't* suited to linear discussions. *But* the fact that everything within a MediaWiki wiki is
- well, within the wiki - has it's advantages too, and I (and, I
suspect, many others) would be wary of just plain ripping out the discussion pages and putting a forum in instead. Better to improve what we have than to replace one set of flaws with another.
The forum structure could be like this:
<snip>
This is where your thinking is already diverging from mine - wikis don't have a hierarchical structure of this sort, and for good reason. One of the longest-standing concepts of a wiki is that, because everything can be editted, everything can be "refactored" - reorganised, renamed, disassembled and put together in a different order, summarised, combined, split...
Although MediaWiki's use of discussion pages has already diverged from that original concept (where discussions are simply "turned into" an article about the subject under discussion), there is still a definite feeling that things are "fluid". And a discussion page on MediaWiki is still just that - a page full of discussion, which can be linked to or not by any related page you feel like, but doesn't need to have a "place in the universe" just *in order to exist*. [Hierarchies are useful, but why should there have to be any one definitive hierarchy?]
Of course, you then have a section of the hierarchy labelled "subforum for each article" - since the articles aren't hierarchical, this has to be one big flat structure; so why do we not just have one big flat structure for the whole site? [BTW, I'm not sure what you are referring to by "portals" here, but that's a minor point]
I think this way the discussions would be much more effective and purposeful.
I agree with the "effective" bit, in as much as they'd be using tools more suited to discussions rather than anything else. I'm not so sure it would make them any less "purposeful" though, since people seem quite happy to define the purpose of a conversation under the current system.
You are here: [Wikipedia Forum] -> [Articles] -> [Foobar] -> [Delete this article?]
Are the first two steps of that example really useful? Effectively they tell you "you are in Wikipedia" and "this is the discussion of an article", both of which are clear from all sorts of explicit and implicit parts of the current design [and if they aren't clear enough, they can be]. This may seem like nit-picking, but what I'm getting at is that the remaining two are also available in the current system, in the sense that the page Talk:Foobar contains all the discussion about "Foobar", and everything under the heading "Delete this article?" can be editted as a section, referenced directly ("[[Talk:Foobar#Delete this article?]]"), etc.
The most important thing is that the discussion is really organized in forums, subforums, threads and posts, EXACTLY the way it is in the known Burning Boards.
The thread and post model is certainly useful. But as for the forum and subforum one, I'm far from convinced. Thus, I can't agree with your "EXACTLY".
Because currently you hardly find ever a real discussion, i.e. about feature requests. Many people out there have fantastic ideas how to improve wikipedia ... so a real discussion forum would be great.
Well, for feature requests about the *software*, we have an installation of BugZilla, because the developers find it useful for tracking things. For ways of improving Wikipedia, the best place is probably the "Village pump", or its equivalent in other languages and projects; that has plenty of discussions on it. Or, of course, you can take advantage of this being a wiki and *do* the thing you're suggesting; then people can discuss its merits with you if they feel the need.
Or when somebody just doesnt understand a part of an article, he could just start a new thread in the article's forum and get an answer...
They can do that already: click "discussion", click "edit" or "+", type your question, click "save". That's not to say we couldn't do with better features for tracking the answers (as a simple one, how about seperating watchlists by namespace? or, more radically, how about giving "sections" enough real existence that the software can answer questions like "when was the last edit *to this section*?"?) and all sorts of other things, but to imply you need a forum before you can ask a question is just wrong.
Like I say, I don't want to dispute the fact that there are definite advantages of forums/bulletin boards which a page on a wiki really can't emulate; but I don't agree that *structure* is one of them. In fact, I would argue that a wiki can give *better* structure in some cases, simply because it is not rigid, and you can create new kinds of relationships between pages and sections just by describing them.
"Eloquence" has described a kind of "ideal world" discussion system at http://meta.wikimedia.org/wiki/LiquidThreads, and I think I agree with the advantages stated there: basically, a threaded discussion gives a solid meaning to "discussions" and "comments", allowing information about those to be processed and manipulated; and, crucially, allowing them to be "watched".
The exact balance between solid relationships and complete editability is a difficult one to get right, and LiquidThreads certainly *sound* good, but might be very hard to get right in practise. One of the key problems, for instance, is how to avoid the system becoming so complex that people are put off taking part in discussions because they don't understand the proliferation of options, views, controls, special cases, etc, with which they are presented.
So my opinion is that wikis can definitely learn something from forums, but I think forums could well learn something from wikis too.
[This protracted ramble brought to you by my tired and distracted brain. We apologise for any boredom caused.]
I read the infos about LiquidThreads, and i've to say: it's IMHO simply perfect. If i'd known about this project, I wouldn't have started this discussion, because at the moment I think LiquidThreads is still *much* better than my proposal ;) So, I think this discussion has not to be continued. If i read correctly, a release of LiquidThreads will only be available with the release of MediaWiki 2.0. I read about MediaWiki 2.0 the first time a few days ago, is there already a code available? Who is mainly responsible for this project? Where are more infos available?
I have dealt with the MediaWiki software for a few weeks now, and i've to say, it's partly really very unlogically organized code ;) So IMHO MediaWiki 2.0 is a real great project. This time, the main target should be - from the beginning on - perfect extensibility, logically structured code (1 parser class in 1 file ;) ), increase of performance and so on ..
[Frando]