Fixing XML

A week or two ago, I was reading something which included a really silly statement hyperlinked to the Wikipedia entry for XML. I followed the link and discovered that the entry was appallingly bad. I looked with a shudder at the size and complexity of the brokenness and just failed to convince myself that it was somebody else’s problem. So we fixed it.

If you want to get a feel for the problem, here’s the August 4th version, as it stood before this work started.

How It Went · As a first step, I sent an email to the old xml-dev mailing list; a remarkable institution that, it’s been in continuous operation (I think) since before XML 1.0 was actually finished in 1998.

Then I started a linear march through the entry, throwing out three or four paragraphs for every one I put in. By the time I got through the first pass, several others were involved, notably including Michael Kay, Rick Jelliffe, Ken Sall, and James Clark. Some of the discussion took place over on xml-dev, but quite a bit is where it really should be, the XML entry’s discussion page.

A Lesson · Right now I’m feeling good about the way this is coming out. Historically, the community of XML experts hadn’t really paid attention to the Wikipedia entry until it was “too late”, when it had become bloated, disorganized, and a theater for pro/anti-XML edit wars. The evidence suggested that the majority of people editing the entry were not overburdened with real XML expertise. Nobody really had the heart to take this mess on.

What changed was this: one engaged party (me) decided it was worthwhile investing the (single-digit number of) hours for an initial hosing-out of the of this particular Augean stable, and knew where to go to appeal for help going forward.

At the moment, my bet is that enough people with an intersection of XML expertise and Wikipedia-editing skills are paying attention that the entry should be in pretty good shape for the next little while.

A Problem? · Well, perhaps. If you go to the XML entry Discussion page, there’s a notice across the top with a big Attention glyph. It says:

An individual covered by or significantly related to this article has edited Wikipedia as TimBray (talk · contribs). This user's editing has included this article. Readers are encouraged to review Wikipedia:Autobiography for information concerning autobiographical articles on Wikipedia.

Well, yep, and the potential problem is obvious. XML is probably going to be the biggest thing on my gravestone after my name. The incentive for me to pump this entry up and make XML seem positively epochal in its importance is huge.

A more subtle but even more pernicious incentive would be for me, while editing the entry, to inflate my importance in the development of XML.

Maybe this isn’t just abstract. At one point a few years back, some XML-haters descended on the XML entry and added a section explaining why it was a crock of shit and anyone with any taste would use YAML or S-expressions or something.

Others who objected, but didn’t want simply to erase others’ edits, added countervailing evidence, and the entry ended up with a section entitled “Criticism of XML” with “Pro” and “Con” lists; sprawling, disorganized, of questionable relevance, and a frequent locus for edit wars.

I raised the question of whether this thing deserved to be in the entry at all; someone took this as Will no one rid me of this turbulent priest? and shitcanned the whole section. I hadn’t been quite ready to do this, but on the other hand I hear no voices raised asking it to be put back.

It would be reasonable to suspect someone like me of maliciously conspiring to censor criticism of my intellectual baby right out of Wikipedia. I honestly don’t know what the right answer is. We want experts to edit Wikipedia, but experts tend also to be partisans. I think the current approach (highlight when it happens) is a reasonable compromise.

Unfinished · The one thing I’m not claiming is that this entry is finished or perfect or not in need of further work. The entry could be better. It should be better. Can you help?

---

---

Contributions