Breaking Things

I had a discussion on Twitter this morning about cataloguing, and the consequences of devaluing and unseeing the labour of cataloguers and other metadata experts. Later I thought about how this relates to the mantra of Big Tech via Mark Zuckerberg: Move fast and break things. Whilst some code4lib and other GLAM tech types like to respond with talk about 'moving slow and maintaining things', the former mantra has definitely been embraced by some of the more careless library managers. Well-structured, thoughtful metadata is difficult to create and hard to explain. It's much easier to see why the labour of cataloguers is important when they stop performing it.

That conversation was sparked by Alissa's passion for metadata, but passion isn't always a positive thing. Travis Kalanick has a passion for illegal and immoral business practices - what he calls 'disrupting the taxi industry' is simply sham contracting and straight-up flouting of local laws. Kalanick is the poster-boy for arrogant and ignorant 'disrupters', but there are plenty of other examples. The common thread is that none of them learned the basic principle that one must understand the rules before breaking them.

So whilst I've certainly advocated for the burn it all down approach more than once, the caveat is that you have to build something even better out of the ashes. It's good to have a passion for simplifying things. It's a useful principle of engineering to say that "perfection is finally attained not when there is nothing more to add, but when there is no longer anything left to take away" - but the trick is knowing when you have reached that point. One person's useless extravagance is another person's crucial lifeline. One only needs to look at the recent controversy over banning single-use plastic straws to see this. Knowing when there is nothing left to take away is really what makes someone an expert. I think about this a lot when slugging my way through a coding project, my verbose, clunky JavaScript standing in stark contrast with the clean, terse code I see written by people who get paid to do this every day. My code usually works eventually, but I'm not yet at the point where I can identify everything that can be taken away - and I probably never will be. It's also why anthropologists are increasingly finding work in tech companies - though usually for the opposite reason, to find things that companies need to add to make their products usable, or safe, or just not super offensive and offputting to large cohorts of potential customers and users.

You can move fast. You can move slow. But just because you're moving doesn't mean what you're doing is progress. When your last metadata expert has retired and you finally just move to automated downloads from a union catalogue, don't be surprised if you've broken a few more thing that you expected.


Building our own house