Jason (jcreed) wrote,
Jason
jcreed

Reading
http://exple.tive.org/blarg/2013/10/22/citation-needed/
and mulling over the idea expressed therein that it's a huge tragedy that We've Forgotten So Much of Our Past already in computer science.

I'm not sure I totally agree, or at least I enjoy trying to construct the counterargument that everything is more or less fine. That it's working as expected, if not as desired.

Tangentially, I do agree that he's correct to observe (and maybe even lament a bit) how easily a plausible myth can outcompete the truth in memetic evolution --- but what I'm talking about is slightly different, namely the idea that it's sad that we've lost good research ideas.

I think at worst it's kind of a bummer, but not a major tragedy. Why did those ideas get forgotten in the first place? At the time of their discovery, their intrinsic effectiveness didn't overcome whatever obstacles of communication or implementation or marketing that were required for them to take off and memorably take hold in the minds of many. I think even in a research community that is doing as good as humanly possible a job of sifting the good work from the bad, some good ideas will be misjudged as bad. You're gonna make some type I errors, you're gonna make some type II errors. And if those ideas really were good, later research might well stumble on them, too.

And so if you operate with a selection bias of "all the good ideas I've ever heard of living in 2013" I think you should expect to find hits among "research papers in 1970" that sound surprisingly prescient. I don't know at what rate I think you should expect to find them, but I expect it to be above zero --- even for a research community doing its best to try to remember the good stuff and forget what didn't work.

That's the crux of my feeling: the fact that you have to choose what to remember. Remembering isn't free. Especially when it's institutional knowledge that has to be actively taught to the next generation. It's hard to teach, and it's hard to judge accurately what should be taught. We try, we win some, we lose some. I'm skeptical of growing too fetishistic about the first person to Officially Discover or Invent something. Rewarding primacy provides an incentive, so it's not completely crazy, but I do feel like the best ideas are those that all but force us to discover them, repeatedly, independently.

(Despite all that, I am totally excited by people spelunking into the past and finding neat old forgotten things. Surely there are plenty of ideas that are potentially great but discovered too early, before surrounding context made them fully realizable)
Tags: research, web
Subscribe
  • Post a new comment

    Error

    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded 

  • 1 comment