Krzysztof P. Jasiutowicz/Metadata pattern talk

A few questions:

• Who will maintain the data?
• Is it practical to expect Wikipedians to maintain the data?
• If it is impractical to expect Wikipedians to maintain the data, how can you guarantee that the data will be accurate?
• When was I appointed editor? :-) What does "editor" mean in this context?
• How can we keep track of who the contributors are? What does "contributor" mean? If I add a comma to an article, am I a contributor?
• How do we decide what "category" something is in, and what the "super" and "sub" categories are?

--Larry Sanger

---

• Who will maintain the data?

The author(s) practicly anyone should have the possibility but the author(s) know the subject best. The other option is to let only editors/administrators do it.

• Is it practical to expect Wikipedians to maintain the data?

Let's wait and see. Rhetorical question - will it be proper to leave Wikipedia without metadata (+additional software to complement it).

I think it's practically self-evident that no one will add metadata except perhaps to your articles, until they're convinced that your scheme is optimal. Why think your scheme is optimal?

• If it is impractical to expect Wikipedians to maintain the data, how can you guarantee that the data will be accurate?

Hey. Don't you see how smart Wikipedians are ;-) If they are so smart to write brilliant articles, let them do some more lines of metadata and in time it will pay off.

Well, my point is that, unless Wikipedians are motivated to maintain the data, the data will be worse than useless.

• How can we keep track of who the contributors are? What does "contributor" mean? If I add a comma to an article, am I a contributor?

Anyone can be. See below.

See my reply below. I'm strongly opposed to all but an automatically-generated list of people who have edited a given article.

• How do we decide what "category" something is in, and what the "super" and "sub" categories are?

Everything belongs to some category. If not it belongs to category 'Other'. For explanation of 'Super' and 'Sub' see my 'Metadata' subapages. Super is more general and sub more detailed. They can be empty as well.

The problem is that everything belongs to multiple categories, and that there are multiple overall schemes of organizing things. As one of Wikipedia's main administrators, and knowing how Jimbo feels about this, I can assure you that we are not going to have any very detailed, complex "hard-wired" category scheme, if we have any category scheme at all. Please see this article if you haven't already.

Kpjas

LMS

---

While I like the idea of having metadata in general, I stongly oppose two uses here: one, any information that can be automatically derived from the software should be, specifically dates of creation, editing, etc. Having humans type things like that is just asking for bad data, and bad data is 100 times worse than no data. Secondly, I don't like the idea of tracking contributors/editors at all. Wikipedia isn't about egos, it's about the product. Articles here don't have "authors" nor should they. Tracking contributors, it seems to me, can only lead to articles being thought of more as proprietary rather than collaborative. --LDC

Software lags behind. You are right that it is much better handled by some software but... Why anonymity in Wikipedia ? We shouldn't force it, should we ? Otherwise we should stick with IPs only. Really good articles are by people that can be identified. I don't say you have to supply your real name. In my opinion collaboration doesn't exclude some form of identification. --Kpjas

I don't totally disagree; if the information is accurate, there's no reason not to include it. I would request that we make it an important matter of etiquette to never write anyone's name in the metadata but your own. To do so is to make a claim on someone else's behalf which they may disagree with. Did Larry really edit the article, or did you just use his name as a placeholder? If someone wants his name there, let him put it there. --LDC

I also like the idea of metadata. It just has to be done right. I think Krzysztof's got a first attempt, but not a final product. I think you've got to work on it more, Krzysztof. And you need to consult other people about it, and persuade them, or else your work will be pointless.

I totally agree with not tracking contributors and editors on Wikipedia. I think we've seen ample evidence that what makes Wikipedia work--so much fun and so active--is the fact that many different people feel they have something invested in particular articles they've worked on and also in the project as a whole. No one is talking about forcing anonymity; there's a difference between forcing and encouraging. I also disagree with the claim that we should stick with IPs otherwise. We are a community, and as such we need to be able to identify each other. Just because we can identify each other, it doesn't follow that we need to "own" our respective main contributions. Again, to do so is to discourage other people from working on them. I suspect, Krzysztof, that you haven't been on Wikipedia long enough to be able to fully appreciate this.

I also stand 100% behind Lee's view that if some data can be generated automatically by software, then please, please let's not have fallible, inconsistent people try to maintain that data. --LMS

---

I will keep adding 'Metadata' to articles I'm the main contributor.
I think it won't hurt the project in any way and it might eventually turn out to be helpful in the future. If not, forget it.
I believe that 95% of contributors will be able to supply some accurate metadata. However to run smoothly, effortlessly and error-free (as much as posssible) - it requires some changes to usemod software or additional software. Altogether different form of editing page driven by other script ?? - why not some other CGI scripts with read-only access to Wikipedia data and writing metadata to a database.
I wonder what server people think ???
BTW I feel that we too often tend to look at Wikipedia from authors' perspective.
--Kpjas

---

First, I don't work for Bomis, Wikipedia, or related companies (or anyone else right now, for that matter—someone make me an offer (grin)), so this is strictly a grunt user opinion. Second, the comments I made in /Proposed keyword search mechanism (about '$METADATA') were more of a "if you're gonna do it, this is what I'd prefer" type comment. So, that said, I think this is too much structure, too soon. Right now, the appeal of Wikipedia is its simplicity. I can just type—as little or as much as I want—and that's it. Other than learning the formatting conventions, I don't have to worry too much about whether I'm using the right "keywords", or pondering which category(s) something should be in. While I think most contributors could supply accurate metadata, I think some would be discouraged by the, perhaps implicit, obligation to do so. I've seen other projects get bogged down in the details like this to their detriment. The more I consider it, the more inclined I am toward keeping Wikipedia simple and structure free for as long as practical. The content and community are still evolving, and it's just becoming useful (cf Larry S's comments in Wikipedia Announcements); I don't think there's an urgent need to rein it in. --loh (2001-06-27)