The WikiWikiWeb

Posted 12 Apr 2000 at 09:23 UTC by mbp Share This

As I write this, I think that it would be nice to have a way to morph diary entries into articles. In fact, that's as good a way to introduce WikiWikiWeb as any other: Wiki makes no distinction between different types of content. Everything is a page, and there is a simple non-HTML markup system. Anybody can edit anything: there is no protection except for a highly-developed sense of politeness and respect.

The most interesting aspect of the mechanics is that reversible links are created by typing WordsSmashedTogetherLikeSo, though links to web sites, books, and email addresses can also be created automatically. This encourages people to automatically string phrases together, and so Wiki seems to discover meaningful hyperlinks by itself.

The system is imbued with a certain austere and ugly beauty. There was a semi-serious adoption of zen/tao attitudes. This acts as a filter on the users: until recently, they have seemed to have a similar outlook.

At least in the past, Wiki has worked devastatingly well, producing content largely on issues of software design, but, like Advogato, branching into the contributors' personal interests.

Wiki is famous as the birthplace of the ExtremeProgramming movement, which itself consciously focuses on simplicity and proper attitude. Wiki was also home to the PortlandPatternRepository, and is a prime force behind the wider understanding and appreciation of the QualityWithoutaName.

Sadly, Wiki seems at the moment to be in the end phase of the CommunityLifeCycle: there is a deal of disharmony, and almost fighting. We have gone from humorous and good-nature joshing to straight-out arguments.

Earlier this year, one contributor felt it necessary to commit WikiSuicide, destroying all his contributions and others beside. The feeling now is somewhat acrimonious and unfriendly. I don't know if it will recover, or indeed evolve to something new and better.

Wiki was a wonderful thing.

WikiWiki, posted 12 Apr 2000 at 11:07 UTC by caolan » (Master)

I read through some of the Extreme Programming stuff previously but I found the extreme hyperlinking such that it was impossible to piece together any meaningful content. Hyperlinking is a great thing, but its not a goal of itself, we could hyperlink every word in every sentence to some sort of related other information, even if it is just a interesting emtymology of the word in question, but that doesn't really create any added value.

What I found the problem with the Wiki.. stuff ,though what I accept is probably what others find the interesting part, is that it is too anarchaic, too disorganized. Each text snippet is too short and small for me, navigation requires a phenominal memory. Its strikes me as like the web would have been if hypertext had gone down the "info" route.

The real problem is probably that it just wasn't designed for me, I browsed it once of twice but you know I couldn't really get a clear image of the purpose of the site. If id grown with the site from some earlier time perhaps I'd have a lot of shared context with the existing user base, as it is I do not. The timeframe for actions mentioned here and there is unclear, etc etc. I imagine that this is one of those things you either love or hate. To my mind linear text, linear flow are good things, and I don't like excessive relience on external information and links, a link-luddite if you will


extreme programming, posted 12 Apr 2000 at 14:19 UTC by Skud » (Master)

Hrm. Extreme Programming looks kind of interesting... it seems to turn a lot of stuff that Netizen thinks about into an actual methodology.

Either DeMarco and Lister (in _Peopleware_) or Yourdon (in _Death March_) commented on the difference between Methodologies (shelves full of close-packed deterministic algorithms for people to follow, as if they had no free will or good ideas of their own) and methodologies (how you do stuff). I have to agree. I'm just not sure whether XP counts as a small-m-methodology (because it's "lightweight") or as a Big-M-Methodology-pretending-to-be-small (because it has its ideas pinned down on paper (ok, on a website) and has developed its own jargon and stuff).

Linking in Circles, Running the Maze, Barefoot in the Tumble-dryer, posted 12 Apr 2000 at 15:36 UTC by Ankh » (Master)

On WikiWikiWikiWikiWikiLostWikiWikiWhereAreWeWiki
I use an autolinking technique too, have done for years, but on phrases that still have the spaces in them, and with an imposed structure. I don't want to get lost. ) Two simple examples are a collaborative IRC Glossary (you won't be able to edit entries without an account) and a 1736 Dictionary of Thieving Slang (canting). The entry for Gypsies has quite a few links in it, all of which (even the See Also ones) are inserted automatically.

For a dictionary or glossary, this sort of linking works well, because the reader is often quite motivated to explore. For serious discussion, I'd prefer to see a constant main article with annotations that move and shimmer in the haze of peer-review. This is part of what Tim B-L was thinking of when he designed the REL and REV attributes to HTML links: you could say that you agreed or disagreed with an article, and he wanted that to appear beside the article itself. Murray Maloney and I published an internet draft for REL/REV once (1995?) and Tim's original paper from 1989 on web architecture is at w3c somewhere if you're interested.

Extreme Programming
There's a lot of good sense there. When I see people trying to manage a large project by running around frantically waving Microsoft Project printouts, I know there's going to be trouble. It's nothing to do with Microsoft, though; the whole idea of running a programming project based on milestones and things that have been done turns out to be useless for prediction.

Programming in the large has been covered extensively in literature, but really, almost all programming is programming inthe small. There's a revent CACM article on this, in the issue with perceptual interfaces on the cover, that's worth reading especially if you are in a large project environment.

Not everything in the WikiWeb about ExtremeProgramming Will WorkForEveryone. Since programming is still mostly an art, not engineering, you have to be willing to adapt.

Extreme Programming [Mm]ethodology, posted 12 Apr 2000 at 16:54 UTC by dan » (Master)

The attraction of XP, I think, is that it looks like a Methodology and behaves more like a methodology. The formalisation of good ideas into Standard Practice is reassuring for people who can't already see that they're good (customers, generally)

Although "Extreme" was a bad choice of name for such people anyway. I suppose you can claim you're doing the Unified Process instead; much of the customer-facing aspect of it is similar

Disclaimer: this is only my recollection on the basis of what I've read. People who actually do this stuff know more them me, Im sure.

Automatic links, posted 12 Apr 2000 at 19:03 UTC by Slow » (Master)

I was actually thinking about this the other day while reading through Advogato. I was thinking of something much simpler however. Just an easy way to link to people and projects listed here. It's not too horrible to have to insert the href manually now, but it would be nice if there were a quick way to link things.
I was thinking, as a practical solution, you could provide something along the lines of <person> and <project> tags for use in comments. My first thought was automatic highlighting, but I don't think that's feasable. For instance, my nick, Slow, and project Entity could both be mislinked in a post.

WiKiWikiWeb is a useful direction., posted 12 Apr 2000 at 20:45 UTC by argent » (Master)

It does have a lot of the feel I was getting at in my "More Links" message.

What it's missing:

  • HTML content (or even UBB markup). The hacks they have to do markup are really clumsy and counterintuitive and undocumented.
  • Ownership of messages. We don't want a WiKiMindWipe.
  • Tracking... there's no backlinking like there is in CritLink.

What's good:

  • Automatic links and link creation. You can link to stuff, and see what's linking to it.
  • You can edit everything you created. Your stuff can be updated.
  • Everything you create has a name. You can find stuff.

What's bad, but not necessarily missing:

  • It's ugly. Even in Lynx.

End of life cycle? The coming of the ZWikis!, posted 14 Apr 2000 at 00:39 UTC by faassen » (Master)

I think wikis are alive and well. While it may be true the original wikis have some difficulties (I've never studied them in detail), wikis are popping up all over the Zope community, in the guise of ZWikis. They are used to discuss a variety of Zope topics, mostly as communal design scratchpads. The advantage of wikis is that they're fairly fluid but still a bit more structured and permanent than mailing list messages.

See the Zope Wiki Central for a listing of lots these baby wikis. There are quite a few by now -- it's very easy to create a new zwiki in Zope, just like it's very easy to install your Slashdot-clone with Squishdot.



Oh, and a reply to Argent, posted 14 Apr 2000 at 00:47 UTC by faassen » (Master)

Oh, and the ZWiki variety offers:

  • HTML content (though most ZWikis have this disabled in favor of Zope's structured-text, which though it has some nits is quite a neat way to create HTML documents from readable plaintext).
  • While I haven't seen ownership of messages (shouldn't be hard to implement tho), most zwikis on need a login before you can edit.
  • Another feature is that Zope contains unlimited undo, so you can undo any damage to a ZWiki.

Disadvantage of allowing HTML is that wiki editing becomes a lot less nice; you suddenly find yourself facing arbitrary HTML instead of plaintext. That's against the spirit of wiki. Too much ownership may also be a bad idea; it'll limit the fluidity that makes wikis so nice. The main idea behind wiki is that they're so easy to change and extend!

-- Martijn

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

Share this page