tom thinks

Humans and Hope, 2001_05_04:14:26

Software development is in progress. Thanks for your patience, or amusement, as the case may be.

home ~*~ index
Metaphysics
respond
responses
When discussing the Edges paper on the Analytic list last summer, Bryan Register kept accusing me of being a Transcendental Idealist, and I failed to explain adequately why I am not, despite often sounding like one.

Now that the Conceptualist Interpretation of Quantum Mechanics paper is out, I feel safe in announcing my metaphysical orientation.

I'm an Empirical Idealist.

That is, some of the things that I say look, falsely, like traditional idealism, but I justify them by reference to empirical observations, notably experimental violations of Bell's inequalities.

That this is a consistent position should been seen as a real problem for traditional idealists, who insist for some reason that we can't know anything at all about reality as it "really is" unless we can know everything about it. My contention is that we can know lots about reality, but there are certain modes of being that are inaccessible to us. So?

This makes me an "idealist" in the eyes of some, who insist, wrongly, that if we don't know all of reality transparently then we must know nothing but the contents of our mental states. So I'm going to keep this label for a while, because it'll raise immediate questions with anyone who knows what the terms mean, and hopefully give me opportunities to break down some of the weird assumptions that are floating around.

One problem that's involved here is simply that the names are so dumb: if idealism just means "not empiricism" then it should be called something like "anempiricism", just as nominalism should be called "arealism". Both "idealism" and "nominalism" are positive terms that carry a huge amount of baggage with them in terms of their implications, none of which is justified simply by virtue of being "not empiricism" or "not realism." The same problem holds with the purported mutual exclusivity and joint exhaustiveness of deontological and consequentialist ethics. They might be jointly exhaustive, but they are certainly not mutually exclusive, as any attempt to even clearly identify any particular ethical system as one or the other shows.

So I'll remain an empirical idealist for as long as it takes for people to agree that they name particular, narrowly defined historical positions that in no way exhaust the possible range of beliefs that one might reasonably hold.
Software
respond
responses
There's always a tension between adding new stuff and cleaning up old stuff. As you add new features you sometimes find that early choices turn out not to be optimal. It used to be recommended that you fix the old stuff now and then go on with adding the new stuff. That's a recipe for trouble, because you're suddenly trying to keep track of a lot more stuff in your head at once. The justification for the recommendation was that no one was disciplined enough to actually go back and fix things once the new functionality was in. This is just false: lots of people are, especially if they work in an organization that values refactoring and has an explicit place in the process for it.

Lines of code is not a very good metric for most purposes, but one area where it is without peer is in evaluating the project life-cycle. You expect kLOC (thousands of lines of code) to go up fairly steadily, but there are key times when the curve should flatten out. There are two ways it can go just before milestones: it can suddenly increase, as late developers code-like-hell to squeeze that last bit of functionality in, or it can flatten out as on-time developers test and massage and tweak their code to iron out the last glitches before daring QA to find a problem. Just prior to release these phenomena are even more distinct: in a well-managed project kLOC will typically decrease at the end of the development cycle as developers find increasingly efficient ways to do things, and cut out poor-quality, repetitive code that was put in to speed up early development.

I've never seen this mentioned in discussions of kLOC as a metric, which is kind of odd. What you mostly see is long screeds explaining why kLOC is a bad metric, and it is if used stupidly, but then McCabe complexity is a bad metric if used stupidly too. Anything is a bad metric if used stupidly--the whole point of metrics is that they're tools, and like any tool can be used badly or well. kLOC is a bit easier to use badly than some, but not that much.
Humans
respond
responses
There's a line from, I think, Wrong is Right, in which one character says to another, "Make sure it happens on camera. If it doesn't happen on camera, it doesn't happen at all."

Most human beings are not very comfortable thinking for themselves, and so they look for sources of legitimacy in their ideas and actions that have nothing to do with the content of those ideas or actions. In the scientific world, a bad paper published in a good journal will be more influential than a good paper published in a less-prestigious journal. This is insane, but it's the way human beings are, so we have to live with it somehow.

I think human communities can be defined by the sources of extraneous legitimacy their members look to. Religious people of various creeds look to the Pope or Immam or what-have-you to legitimate their ideas. Socialists look to a certain fragment of the academic literature as well as particular media outlets. The "mainstream" looks to the New York Times and the major broadcasters, and the "journals of record" in the various scientific disciplines.

Then there are the "alternative" communities, who look to the literature produced by their own kind. Such literature lacks the kind of self-critical aspect that informs the best of the mainstream communities--members of alternative communities are typically looking for validation and confirmation of their own experiences and beliefs, not critical reconsideration. Thus, the notion of a "dissenting objectivist" is practically an oxymoron, or as would be a "scientology heretic". To be a member of a community means the uncritical acceptance of something, and alternative communities tend to be a lot narrower in what they expect you to accept.

Those of us who rely primarily on our own minds don't belong in any of these communities--we're out drifting around on our own, trying to make sense of a world that contains nothing we take for granted, nothing that we can simply assume without thought or consideration. We accept the evidence of our senses, but not blindly, not by simply assuming but after arguing for and against it and coming to see how consistency requires us to accept it.

Sometimes, if we're very lucky, we meet someone who shares some of our own conclusions, who has reasoned the same way we have, perhaps made the same mistakes, and followed the same twining paths of logic and guesswork to make a kind of sense of the world. In the best of all possible worlds, we meet someone who has a viewpoint that is just different enough from ours to be a source of constant education and delight.

That's the world I live in.
Notify me when tom writes again. (You may subscribe now, but mailing will pause during testing.)

Find Enlightenment