The sad evolution of wikis | Hacker News Reader

The sad evolution of wikis

apenwarr.ca

91 pointsblasdel44 comments

I've been lurking on the c2 wiki (and occasionally posting, http://c2.com/cgi/wiki?ScottVokes) for almost ten years, and the culture there feels completely different from Wikipedia. The c2 wiki isn't a blip on most programmers' radars, though, let alone the world at large.

Back when I was preparing to become a librarian (academic research or possibly a medical archivist; before I got sucked into the software engineering vortex), I spent a fair bit of time talking with my coworkers and professors about the relationship between rapidly-mutating secondary (tertiary?) sources, such as Wikipedia and blogs, and actual primary sources based on actual hardcore effing research. The c2 wiki did not once ever come into the discussion, because nobody cared about HowComeLispAndSmalltalkAintPoisonYet or DoesRelationalRequireTypes. It was too alien a culture.

I think the author is on to something, that the Wikipedia culture of anal-retentive editing and (worse yet) knee-jerk "[citation needed]" has been bleeding the spontaneity out of the popular concept of wikis, so people are much less inclined to just post stub documents and collectively flesh them out as time allows. Often, the expectation that writing ends up polished prevents it from appearing in the first place. (Documentation or otherwise.)

I've successfully pushed for a wiki literally everywhere I've worked through the last decade (including the library system, which has a great and actively maintained collection guide, based on mediawiki). Sometimes it took offering/threatening to run my own server specifically to host one, and it's always taken some content-seeding and nudging to get things in motion, but the net results have always been positive.

At the last place I worked there was some use of Sharepoint, but we were forbidden from using it for documentation after it ate a ton of our sysadmin's documentation (the only guy who was using it, that is.)

And of course this being a Windows shop the idea of using something else was too weird. "We do Windows."

We had a sharepoint-based wiki at one point. We salvaged as much content as we could and moved it to mediawiki, which actually worked* . Half our office was convinced sharepoint was deliberately designed for incompatibility with firefox. (Is there such thing as a "cygwin shop", by the way?)

* Usemod is also quite good, but less stylish.

Pretty sure it's just the normal "works only with IE" stance of most Microsoft stuff. I did some testing with Firefox and Sharepoint for work to see what doesn't work and it's fairly extensive. The worst issue stopping non-techies is that the rich text editor for textareas doesn't work in FF, you have to type in raw HTML. Pretty much a showstopper. This was SP 2007 btw, not sure if 2010 is any better. The blog and wiki features in SP2007 felt tacked on and incomplete to me anyway.

The thing that makes Sharepoint completely unusable for me is that all the links are of the javascript() variety, so quickly opening pages in tabs is functionally impossible.

It would be nice if browsers could handle middle clicks a little more gracefully in this situation.

I've found c2 wiki very interesting, mainly for XP related stuff (eg criticism of the Visitor pattern; do the simplest thing that could possibly work) - but it's not a public wiki, because you can't contribute without a password, which only the select have (no idea how they got it.)

Reading it reminds me of a mailing list archive of intelligent discussion. Haphazard, sequential, with links only at times. There are very insightful thoughts tucked away in there, but it takes a lot of reading and tracing to find them - and then you can't find them again without going through the same process. It's abysmal, labyrinthal, like some of my prototype projects.

You can't contribute without a password? Every time I've been there, they've had the world's most half-assed captcha ("the password is [three-digit number], type it in the box"). I had no idea I was so elite.

There's a lot of great content, buried in a lot of random rambling. (And some of it is mine, sorry.) It's like HN, but without hueristic fading for old threads, and with a culture focused on software-engineering rather than startups.

This was years ago. I must have had bad luck to have wanted to edit 'special' pages, that don't supply the code. I assumed that was the general rule, which the help page seemed to confirm:

> When no code word is given, you may use an alternative code word that floats along among trusted users. There is no way to apply for this code. Try your friends-of-friends network. Remember that this code and the mechanism itself will change if abused. http://c2.com/cgi/wiki?MoreAboutCodes

I just checked what you said, and the above seems to be (now?) the exception, and most pages do supply the code. A surprise to me, good to know, thanks!

It was probably only enforced for pages that got a lot of vandalism, during heavy slashdotting, etc. It's definitely the exception rather than the rule.

There was times where the digit was written, other times where the digit wasn't. Only some people got the password then by word of mouth.

Ward was really willing to do something that could work for blind people as well and there wasn't really audio captcha at that time.

In my mind it didn't work really well as the community was too large, and as a wiki require a great deal of people to have it maintains such a large wiki it essentially collapsed under it's own weight and word of mouth never arrive to me (I was just a lone french with no connexions )

Not knowing whether you were going to be able to modify or not modify it at a time was also very annoying.

mediawiki it also a very complex wiki, and as it somewhat became the standard, non programmer got afraid by it and didn't want to edit anything by fear of breaking something or not knowing how it work.

I had it explained to me that "MediaWiki's purpose is to power Wikipedia. Your personal wiki is a side effect."

usemod (http://www.usemod.com/cgi-bin/wiki.pl) is at a sweet spot on the features vs. complexity continuum. It's also pretty easy to chroot.

I've also read on the c2 wiki. TVTropes also fits your third paragraph about no expectation of polished writing in stubs.

Yes, and it also fits the category of wikis that tend to cause insomnia. :)

I started MeatballWiki (http://meatballwiki.org) just over 10 years ago to wax poetically about wiki culture, codify it, and try to reproduce it in other environments since I thought it was superior. However, Wiki cultures have been difficult to keep sustainable let alone easily reproducible.

All non-reproducing communities need both openness to attract new people faster than people leave, but closeness to protect itself from negative people. Most wikis lack an ability to close themselves off in a non-hostile manner (i.e. 'turtle mode'). You may laud the wiki culture method of keep order, but I've helped hundreds of wikis defend themselves with "soft security" (http://meatballwiki.org/wiki/SoftSecurity), and I found it is actually a cynical process.

In terms of keeping the culture going, since wiki process is based almost entirely in social custom, it requires a lot of human energy to teach newcomers and fend off attackers. Teaching requires more effort than physical-world communities because of the low bandwidth of communication on a wiki.

I think there might be a new evolution one day where some of the older wiki cultural principles will be reimagined, perhaps unknowingly, in media that has a better balance of human effort to community value. After all, healthy wikis are a wonderful experience so there is no reason they cannot continue to thrive and adapt.

You probably feel like you're yelling into the void of space, but I really appreciate what you've done with Meatball and Usemod. If you ever come through Michigan, I'd be happy to buy you excellent ales, coffee, tea, etc. (Beer is really our strong point, but I also know some amazing coffee roasters, etc.)

To me it sounds like the success of NitWiki was down to the culture of the 30-person team. I'm guessing this stemmed from some leaders in that group, and applied to more than just the wiki. I'm not sure that much has changed over time - it's just not trivial to cultivate the right kind of culture.

Re: the public vs private debate, if you want to make a great public wiki, then maybe you have to make "open" the default. This requires letting go of the fear of sharing too much, although my personal experience suggests we tend to be more afraid of this than is necessary (e.g. "stealth mode" startups).

Before 2005 most Wikis used to be easy to use, considering the target audience at the time (computer literate people).

Nowadays, with 500 millions on FB, the audience has become "the rest of us".

Unless I missed something, I think that there is currently no easy wiki for this audience. Neither the older wikis, c2 style, nor modern ones, wikia or wetpaint style, for example, are simple enough.

I mean, what wiki service would you direct your kids or your parents to?

That's a sad situation. But it is also an opportunity.

I am working on a dead simple wiki, c2 spirited, but 2010 style. This is alpha software at this point. You may want to pay a visit (and, yes, WikiWords work). Sometimes it run at http://simpliwiki.com. Feedbacks welcome :)

Currently down:

    The following error was encountered while trying to
    retrieve the URL: http://simpliwiki.com/

        Connection to 184.72.45.204 failed.

    The system returned: (110) Connection timed out

It is now installed on an Amazon EC2 micro instance, much cheaper than the small instances, as a result I don't stop the instance anymore.

Thanks for trying, maybe another shot?

I believe there are services that allow one to make annotations on web pages. Perhaps they allow one to control the degree of sharing; that might go some of the way toward what Mr. Pennarun wants. An annotation might link into or out of a private wiki.

I'd like something even more modest: a better place than Wikipedia to discuss software. Wikipedia policies just don't allow what I want for current software. E.g., look at the near uselessness of http://en.wikipedia.org/wiki/ReStructuredText (if it's still there?) and then look at the discussion page and history to see all the useful content that's been deleted. Or just read the text next to the broom to feel the essential "Go away!" point of view. Is there a better wiki for software topics? Or are Google searches the best one could hope for?

Try the original wiki (http://c2.com/cgi/wiki?FrontPage)! It lacks the anal "[citation needed]" culture that wikipedia has acquired, and it's been focused on software since its birth. (I've been lurking and occasionally posting there for about a decade.)

They have obvious biases (a lot of the original content was written by Smalltalkers and XP / Agile people, for better and worse), but it's worth checking out.

I used to think Wikipedia is neutral. Lately I've been kind of shocked - I don't know if this is a new development, but check out these two pages:

http://en.wikipedia.org/wiki/Augusto_Pinochet

http://en.wikipedia.org/wiki/Fidel_Castro

Look at the tables of contents. If you didn't know better, you'd think Pinochet was more corrupt and murderous. But, Castro killed 10x as many people, tortured 100x as many, stole orders of magnitude more money and wealth...

There's 36 links in Castro's Wikipedia table of contents. Only four of them are mildly negative: 9 Controversy and criticism, 9.1 Human rights record, 9.2 Allegations of mismanagement, 9.3 Allegations of wealth.

There's 24 items in Pinochet's table of contents. 10 of them are very negative.

2 Military coup of 1973, 2.1 U.S. Backing of the Coup, 3 Military junta, 4.1 Allegations of fascism, 4.2 Suppression of opposition, 5 Arrest and trial in Britain, 7 Secret bank accounts, tax evasion and arms deal, 8 Human rights violations, 9.1 Demonstrations

If you didn't know better, you'd think Pinochet was a very bad man and Castro was a good man. But regardless of your politics, Chile is a much nicer place to live than Cuba these days, yet Cuba was a much nicer place than Chile before Castro came to power. Castro's policies destroyed Cuba. Whereas Pinochet has a mixed, but pretty good record. Hell, Chile just passed net neutrality laws last week.

A similar article is this one -

http://en.wikipedia.org/wiki/The_surge

By almost all analysis and objective metrics, the 2007 troop surge worked. Violence went down, insurgency went down, attacks went down, crime went down afterwards. But if you look at the article, it's loaded with criticisms, opposition, and things unrelated to that particular military campaign. Then there's a quick blurb, "Interpretation of the surge's results ... Whether the surge led to the improvement in Iraqi security, or other factors caused it, is disputed by some."

Wikipedia is not neutral. It's edited by people who have more free time on their hands are and are more tech savvy. People with more free time who are more tech savvy tend to skew demographically, which is sad, it should be a neutral resource and would be better if it is.

Compare the Castro and Pinochet articles if you want to see it in action, you'd never realize that Pinochet was a much better and more effective leader. They both did bad stuff, but Pinochet did less bad stuff and more good stuff. His country was worse when he took power, and better afterwards. Cuba was much wealthier, prosperous, and better off before Castro, and is much worse now. It's a shame that this is misrepresented in Wikipedia.

Both Castro and Pinochet were military dictators that killed and tortured a lot of people.

To compare how good or bad they were by comparing how prosperous the countries are is silly, one of the reasons being the external influences.

Pinochet's junta "disappeared" over two thousand people and tortured over 30,000 (people with names, not estimates). I don't see any credible source stating that Castro killed 20,000.

I actually see this type of Castro/Pinochet comparison as a test for Human Rights and checking if someone is inconsistent because of their right/left wing preconceptions. I find most people dislike or hate one while justifying more or less the other when they are both basically the same.

> I find most people dislike or hate one while justifying more or less the other when they are both basically the same.

Fair enough, I'm not here to argue the merits of one murderous dictator over another. The point is that Wikipedia has 36 sections on Castro with only 3 slightly negative, and 24 on Pinochet and 10 of them are very negative. You'd agree there's a bias there, yes?

Well, I didn't make a comparison study between the two articles but I just read the first ones on Pinochet that you said were "very negative" (Military coup of 1973 and U.S. Backing of the Coup) and they don't seem negative to me at all, pretty mild if you ask me. This overthrowing of a government would be the equivalent of Castro's revolution so I don't get your point sorry.

First time I've ever defended Castro even partly. :-)

>>This overthrowing of a government would be the equivalent of Castro's revolution

Arguably, Castro overthrew a dictator. Allende was, afaik, quite democratic(?).

Castro did break promises about general elections. He took over and got military/economic support from other dictators.

The main difference is that today Castro+family are still dictators of Cuba -- while Pinochet and his junta are gone since a long time, leaving a quite nice country.

And this is totally off topic for HN and the article, of course.

By almost all analysis and objective metrics, the 2007 troop surge worked. Violence went down, insurgency went down, attacks went down, crime went down afterwards.

The surge was enacted with a particular goal: create breathing space to foster political solutions to Iraq conflicts. It has failed. The political solutions never came about.

Beyond that, claims that the surge caused a decrease in violence require evidence. Correlation is not causation. And there are many other reasons that explain a decrease in violence in the same time frame.

Please don't turn a discussion of how wiki's are no longer the free form web-based collaborative scratch pads they were envisioned to be, into some stupid political nitpicking based on some specific wikipedia article not being "neutral" (in your viewpoint's favor). The point, is that wikipedia has turned the popular idea of a wiki into a formal documentation system, and that idea has gone viral into other places where such clinical sterility of expression is not appropriate.

Using wikis to manage a small team's internal documentation and shit is a PITA, particularly because you get into a situation where you need to sort of replicate some formatting on a bunch of different pages but then, whoops, nobody knows how to use the template system in a sane way.

Live preview and some sort of not-shit WYSIWYG would make wikis for small teams much nicer. Then what'd make it even super nicer is lightweight macros and pre-made templates that could be stored in a per-user or per-category way. Like... github, but for wiki-stuff.

have you tried confluence? Anyway, even in the free/open world, a lot of wiki software come with simple wysiwyg editors and templates

As a Confluence user I have to second that.

For a small team Confluence is an even more "no-brainer" than MediaWiki, etc...

First time I ever heard something positive about confluence. (But I only heard about it, because we use it at work and people like to complain about it.)

FWIW, I don't really like confluence either (I do have an high opinion of other atlassian's products though), it just seems to fit the OP's wanted-list :)

Could you elaborate on what their complaints are?

To me its waaaay better than Sharepoint - which we also used. And it is way more suitable for "non-technical" and "non-wikisavvy" users than e.g. MediaWiki.

I have never used Sharepoint. Fortunately I do not work in a Microsoft shop. (Though we do have to bear Exchange Server.)

One minor misgiving about Confluence is its syntax: Opening and closing tags look the same. So {noformat} both starts and ends its environment. I haven't done anything fancy myself with Confluence, so I can't comment on the other complaints. (And we do use an ancient version, because we have some custom changes, and nobody has found the time to port them to the new version.)

One paragraph stood out:

Unfortunately, outside of wikis, the world itself has also been busy evolving, and in a way I don't know how to deal with. The new problem is: teams are the wrong size now. Thirty dedicated, full-time programmers was an ideal number for NitWiki. But who has 30 dedicated, full-time programmers now? I mean really dedicated? You can be a very successful Internet startup with way less than 30 developers. Maybe you need only two. Maybe those two aren't even working full time on the one project. And two people don't need a wiki.

Has the programming world changed so much in ten years that 30 person teams no longer exist? What do people think?

In a thirty-person team, you have smaller groups working on different things, all nominally coordinated by the manager who in practice heavily depends on capable developers to help manage the subgroups. Since every convenient method that exists in a large company will eventually be formalized and fetishized, it was inevitable that someone would seize on the idea of breaking up the larger team and formally replacing it with a set of small teams with fluid membership.

Now a large team is seen as grotesquely unwieldy, bureaucratic, and high-overhead, kind of a laughable concept compared to smaller teams, even though large teams were normally run as a set of fluid, informally organized smaller teams. With a single large team, a certain amount of mandatory overhead is incurred at the team level, and individual groups within the team can adopt as much or as little overhead as they think appropriate. One person can work alone, or all thirty people can work in concert. Informal groups within a team never have a formal existence, so they can form and unform at will. Everything depends on judgment and discipline, and the greatest risk is chaos. Sounds kind of lightweight and agile compared to an "agile" process like Scrum, where each small team has a formal existence, a minimum size, mandatory administrative overhead, and a prescribed methodology, doesn't it?

Xanadu? Ted Nelson, not Olivia Newton-John. http://xanadu.net

I think I know what you mean, but could you elaborate? It's a non sequitur otherwise.

(early vision of the internet) not (the movie), presumably.

I was attempting to answer this question: "How do you create a vibrant community, but allow for private topics and discussion, but allow for public topics and discussion, and allow me to work for more than one company at a time with multiple private discussions, and have my WikiWords always end up pointing where they're supposed to?"

Xanadu was a pre-web attempt to create all that the web is and more. It failed because it was too ambitious and too complex technically. For example, links were two way, so from a document you could find all the documents that pointed to it. It supported micro-payments. I think it was going to support itself by skimming from the payments. It was pre-internet, so they had to invent a communications infrastructure on top of POTS. I visited Xanadu for an afternoon in the mid 80's. The programmers were talented and very dedicated, but it was obvious that they were having difficulty converging on something that was simple and worked.

Contrast this with the Web which had two simple protocols, HTML and HTTP, which could be implemented easily and extended by the multitudes.

Reply on news.ycombinator.com