Four stages in the relationship of computer science to other fields

This weekend, Oliver Schneider — an old high-school friend — is visiting me in the UK. He is a computer scientist working on human-computer interaction and was recently appointed as an assistant professor at the Department of Management Sciences, University of Waterloo. Back in high-school, Oliver and I would occasionally sneak out of class and head to the University of Saskatchewan to play counter strike in the campus internet cafe. Now, Oliver builds haptic interfaces that can represent virtually worlds physically so vividly that a blind person can now play a first-person shooter like counter strike. Take a look:

Now, dear reader, can you draw a connecting link between this and the algorithmic biology that I typically blog about on TheEGG?

I would not be able to find such a link. And that is what makes computer science so wonderful. It is an extremely broad discipline that encompasses many areas. I might be reading a paper on evolutionary biology or fixed-point theorems, while Oliver reads a paper on i/o-psychology or how to cut 150 micron-thick glass. Yet we still bring a computational flavour to the fields that we interface with.

A few years ago, Karp’s (2011; Xu & Tu, 2011) wrote a nice piece about the myriad ways in which computer science can interact with other disciplines. He was coming at it from a theorist’s perspective — that is compatible with TheEGG but maybe not as much with Oliver’s work — and the bias shows. But I think that the stages he identified in the relationship between computer science and others fields is still enlightening.

In this post, I want to share how Xu & Tu (2011) summarize Karp’s (2011) four phases of the relationship between computer science and other fields: (1) numerical analysis, (2) computational science, (3) e-Science, and the (4) algorithmic lens. I’ll try to motivate and prototype these stages with some of my own examples.

The first stage is the numerical analysis of X. This is solving the equations that already exist in the field X but are too big for pen-and-paper. A classic example of this for physics would be from the Manhattan Project: MANIAC, IAC and ENIAC, processing for sixty days straight for the engineering calculations required to build the hydrogen bomb.

The second stage is computational science of X. Often this is abbreviated as just computational X. This is when we move from just automating the solution of equations to the sort work we wouldn’t even consider on paper: simulating and visualizing the objects of X. In the case of physics, it can at times be hard to draw the line between numerical analysis and computational sciences but in less mathematized-fields, it is usually much more clear. In biology, for example, it would be running agent-based simulations of evolution. From the visualization side it might involve all the algorithms associated with bioinformatics.

The third stage is e-Science of X. This is the name that I am most ambivalent about. This is when we manage extensive experimental data and method for collaboration over the internet. A biological example might be something like folding-at-home, or the various queriable gene or disease databases. In physics, this might involve historic examples like Tim Berners-Lee developing hypertext to help physicists collaborate on high-energy physics projects and inadvertently giving birth to the world wide web. A more recent example might be all the engineering and computer science involved in getting data out of Large Hadron Collider or in synchronizing the various observatories around the world for the black hole image. More broadly it might involve books like Michael Nielson’s (2011) Reinventing Discovery.

But all three of the stages view computer science primarily as a service provider to field X. As a mean to do better the things that field X would do anyway. They don’t fundamentally change the basic objects and theories of X. The computational science stage much shift emphasis: for example towards token-based rather than type-based models in evolutionary biology. And the discoveries that these resources might facilitate could change the field. But the computer science itself in stage one through three is seldom the direct cause of the change in itself.

This is why I can’t agree when Markowetz (2017) writes that all biology is computational biology, for example. All biology might (or should) use computational biology. But this service role in itself does not place computation at the heart of biology. For that we need algorithmic biology.

The fourth stage is the algorithmic lens on X. Computing as a universal way of thinking. This is when we recognize that theories and other theoretical objects are themselves specifications of objects, and all physical processes can themselves be viewed as computations. Once this link is made, theoretical computer science becomes part of the field itself. It’s theorems and perspectives become parts of the bedrock on the field. This is what a theoretical computer science interested in natural science field X aspires to. Karp, Xu & Tu actually call this stage the ‘computational lens’ and might have a slightly broader (but also less ambitious) view of it than myself. But I think it’s better to have a more distinct name to differentiate algorithmic X from computational X.

Regardless of name, these stages are very theory focused. They put theory on a pedestal at the top of the stages. And, in general, this is an unreasonable view. Do you, dear reader, have some suggestions for better categories? Ones that aren’t linear or hierarchical. Ones that don’t ‘end at theory’?

And if Karp, Xu & Tu’s four stages are reasonable: at what stage is your favourite field? Is this good, bad, or irrelevant?

References

Karp, R. M. (2011). Understanding science through the computational lens. Journal of Computer Science and Technology, 26(4), 569-577.

Markowetz, F. (2017). All biology is computational biology. PLoS Biology, 15(3): e2002050.

Nielsen, M. (2011). Reinventing discovery: the new era of networked science. Princeton University Press.

Xu, Z. W., & Tu, D. D. (2011). Three new concepts of future computer science. Journal of Computer Science and Technology, 26(4), 616-624.

Advertisements

About Artem Kaznatcheev
From the Department of Computer Science at Oxford University and Department of Translational Hematology & Oncology Research at Cleveland Clinic, I marvel at the world through algorithmic lenses. My mind is drawn to evolutionary dynamics, theoretical computer science, mathematical oncology, computational learning theory, and philosophy of science. Previously I was at the Department of Integrated Mathematical Oncology at Moffitt Cancer Center, and the School of Computer Science and Department of Psychology at McGill University. In a past life, I worried about quantum queries at the Institute for Quantum Computing and Department of Combinatorics & Optimization at University of Waterloo and as a visitor to the Centre for Quantum Technologies at National University of Singapore. Meander with me on Google+ and Twitter.

One Response to Four stages in the relationship of computer science to other fields

  1. Pingback: Space-time maps & tracking colony size with OpenCV in Python | Theory, Evolution, and Games Group

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.