Is Computer Science a Science? A challenge for you!

Is Computer Science a science? My mind is set on this debate and I’m not interested in debating it.

Most of my papers have an experimental section, but does it follow I do science? Actually, I spend a lot of time crafting my experiments so it is an integral part of my research. However, most of these experiments follow the reverse scientific method: I seek experimental evidence to support my claims or to quantify more precisely my theoretical results. I’m not alone since I very rarely see papers in Computer Science whose point is to falsify a theory someone promoted. You just don’t see these types of paper. Other times, when I’m not publishing papers, I simply hack: I use technology in beautiful ways. It used to say on this blog that I am a hacker. It still says so in my RSS feeds and it is still true. But I think I’m a hacker in the noble sense of the word since I’ve never hacked into someone’s system. I think the Semantic Web, if it succeeds, will be a giant and beautiful hack. But I don’t think it will be part of a scientific theory.

However, I’m interested in issuing a challenge. Let’s assume that the Semantic Web is a scientific theory or, at least, a set of scientific hypothesis. By the virtues of the scientific method, these hypothesis must be falsifiable. Hence, it must be possible to falsify the Semantic Web. Can you describe a set of experiments testing whether the scientific hypothesis underlying the Semantic Web hold? Once you are done, repeat with related topics like Learning Objects .

Note: Among other things, I’m helping to organize the 2006 Canadian Symposium on Semantic Web and I recently wrote a paper on learning objects so don’t come blasting through my door yet if you are working on these topics and think they are worthy of consideration.

2 thoughts on “Is Computer Science a Science? A challenge for you!”

  1. First off, falsifiability is only one criterion for a science. Not all philosopher’s of science agree that it’s as central as Popper(?) claimed. It’s not obviously useful when talking about social science, where there is not an objective background against which a theory can be complared. The research goals of learning object researchers and semantic web researchers rarely involve understanding nature. They have no external objective standards to compare themselves to; at best, there are internal consistency and coherence standard.

    Social scientists measure things like the opinions, beliefs, reactions, etc. of *people*. The semantic web and learning objects are both apparently interested in satisfying various people (i.e. “web users” and “learners”), and so the only sensible measure of its overall success is a measure of how satisfied these people are. Thus they are clearly in the domain of social sciences. You are not judged by how well you match up with nature; you are judged by your users.

    Practical measures of success include uptake by other scientists (especially in different fields), or use by companies. It strikes me as close in spirit to economics.

  2. Of course, there is a wide range of activity in Computer Science, so it’s probably not possible to say something like, “Computer Science is X,” where X is one thing. Having said that, I wold say that a good part (the most science-like part) of Computer Science is math. CS by and large doesn’t seek to understand the natural world. If we seek to understand anything, it is the properties of formal systems: computers. Some folks just like to play with formalisms. To the extent our work isn’t in the form of formal proofs, its just that we lack the tools to do such proofs.

    The other major part of CS is engineering, which of course is applied science. But again, we’re mostly not trying to understand the natural world, but rather to use our knowledge to build devices (in the broadest sense, including software as a device).

    Having said that, there are those few people who actually do science in CS. Some of the Human Factors people try to understand human perception and cognition in the context of interaction with computers (or not in the contex, even).

    I don’t think this is a matter of hard versus soft science. To me, the distinction there is not the goals but rather the lack of deep theory for the latter. Most of CS has different goals.

Leave a Reply

Your email address will not be published. Required fields are marked *

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    
    Markdown is turned off in code blocks:
     [This is not a link](http://example.com)

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see http://daringfireball.net/projects/markdown/syntax