It’s an entertaining read, Doctorow’s Metacrap : Putting the torch to seven straw-men of the meta-utopia. It’s been 13 years later, and we need to see if this article is a harbinger of the meta-doom, or if something changed to prevent it.
First: note the title. Doctorow admits he’s fighting straw-men. Let’s not pretend that we’re oh-so-smart because we can refute his arguments. Instead, let’s see what we still need to be careful about, and think about how we can better improve on the last few years. I’ll try to go quickly through each of the points:
- People lie: And people still do. Take a look at The Worst Things For Sale if you want to see the things people lie about. But it’s better: You can search without getting just loads of useless links. I was in the trenches of the Web in 2001; Doctorow is being kind when he describes how bad the results were. But the metadata didn’t improve; the software and algorithms used to process it improved. As for the other points (spam, blatantly poor advertising), please recall that we don’t use metadata to record the truth, but information. Information can be false, and we can use other metadata to note how true the information is.
- People are lazy. No change here. Lazy people don’t get their information found unless non-lazy people help them. If you like your aunt’s photos, then take some time to help her scan and tag them. Go on. I’ll wait. Heck, archives help with items from dead people, and the dead are notorious for not helping with metadata.
- People are stupid. Librarians prefer the term “in need of assistance.” Again, a librarian’s response to someone needing assistance is to give it. If we don’t, then we lose our jobs, because all we do is assist people. It also helps us to remember that we all have our days when we “need assistance.”
- Mission: Impossible — know thyself. Even the smart and the willing have problems determining metadata. At the best of times, we’re biased. At the worst, we’re biased and arrogant. Nor can we depend on the kindness of others to balance our biases; remember, “to really mess things up requires a committee.” So, how do we counter our own biases? The hard way, alas. We remain humble about what we can do, and we never consider anything to be the last version. We keep creating, and revisiting, and corroborating, not for perfect metadata, but for pragmatic metadata.
- Schemas aren’t neutral. Okay, we’re moving from the problems with people and moving to the problems with schemas. This goes hand-in-hand with the next one. By organizing information, you privilege some information above others.
You see where Doctorow’s going now, right? The argument isn’t, “Metadata is useless.” It’s “Please stop thinking that you’re going to find this one metadata schema that you can apply to everything you work with and you’ll be done.” It’s an argument for humility in metadata. And goodness, librarians can be arrogant about metadata. Sitting in cataloging class during the lecture on music materials, I did a search on LCC for musical recordings. Turns out, LCC is really awful for classifying musical recordings, but dang, we sure try! We keep faceting this, and shoehorning that, and heading back for rewrites when we discover that LCC can be really Anglo-centric, and we start to wonder what’s wrong. Nothing’s wrong; you just forgot to check your metadata toolbox for the right one for this job. In the example from Doctorow, both washing machine manufacturers are going to be really disappointed when they see the schema made by the guy who just wants to sell washing machines by both of them.
- Metrics influence results. See above. When we see results are off, we find new metrics, and maybe keep the old ones for results comparison.
- There’s more than one way to describe something. Okay, I’ll end on a little slagging on Doctorow. Why all the haters for controlled vocabulary? And what’s this pathetic argument from Doctorow? “Denud[ing] the cognitive landscape”? “Enforc[ing] homogeneity in ideas”? “It’s just not right”? “It’s just not right” is what children say when you point out that they can’t just grab money from the bank in Monopoly! It literally means “I can’t argue this, so I’ll just yell.” Yeah, there’s more than one way to describe something. But how will we know if they’re describing the same thing if we don’t connect them through, say, a metadata structure? And a controlled vocabulary doesn’t mean that we throw away the original description once we’ve tied it in. Metadata does not replace data. I promise you, Doctorow: Next classification schema I’m involved in, I will totally fight for “Solar panel — sex machine usage: See BALD SPOT”.
So, there we have it. We continue to make adjustments to our metadata, our algorithms for using metadata, and the business practices involved in creating metadata, not to reach the metadata nirvana, but to say, for a couple more months, “we met a use for our patrons.”