Mailund on the Internet

There’s a nice short article in American Scientist titled How do scientists really use computers?

An interesting read if you, as I, teach life scientists (and not computer scientists) computer science.

The conclusion doesn’t surprise me much, though:

Our results can be interpreted in many ways, but I think two things are clear. The first is that if funding agencies, vendors and computer science researchers really want to help working scientists do more science, they should invest more in conventional small-scale computing. Big-budget supercomputing projects and e-science grids are more likely to capture magazine covers, but improvements to mundane desktop applications, and to the ways scientists use them, will have more real impact.

Even at BiRC where we do a lot of genome analysis that really do need computer grids, most of our computer use is desktop computers.

–

219-220=-1

This just isn’t right!

How can you feel entitled to a well paying job just because you paid tuition for a degree? I mean, come on!  It baffles the mind!

You pay for tuition, so if you get taught then you get what you pay for!  If you didn’t learn anything, well, that is is a problem, but unless it is the teaching’s fault you have no one to blame but yourself.  You can sleep through the lectures and not do your homework – that is fine by me – but you cannot blame the teacher later on.

Even if that isn’t the case, and you worked really hard and actually learned a lot, how can the university be responsible for you getting a good job down the line?  Once you get the degree, you are on your own!  The job marked might be bad right now – or it might always have been (which probably isn’t the case here since we are talking IT, but it is not quite as attractive now as it was during the .com bubble) – or the competition might just be better, but how is the teaching institution responsible for this?

Yes, they might have given you the wrong impression about the studies, but cave emptor, baby, cave emptor.

I’m pretty sure the university never gave you any employment guarantees.  You have no-one to blame but yourself here.

I feel sorry for both of them, really I do.  No one should be in this situation if it can be avoided, but you just cannot blame anyone but yourself for this one…

–

215-214=+1

There are two interesting papers in the last issue of Science concerning biology education and the need for computer science and mathematics as part of it:

Computing has changed biology – biology education must catch up

Pevzner and Shamir, Science 31 July 2009: Vol. 325. no. 5940, pp. 541 – 542

Mathematical biology education: beyond calculus

Robeva and Lauenbacher, Science 31 July 2009: Vol. 325. no. 5940, pp. 542 – 543

and a great piece at Ars Technica discussing them.

The message in the two papers is that computation and mathematics is such important aspects of modern biology – especially molecular biology – that biologists need to learn more of it and need to fully understand the computational tools they use.

The Ars Technica piece asks the valid question: Why?  To develop computational tools for biological data analysis, of course you need to know, but do you just to use the tools?

Now, I tend to agree that computer science and mathematics should have more focus in biology education – and not just biology but any science really – but I also agree that we need to ask the question why should they learn what they learn? and to which degree?

It is not possible for everyone to be an expert on everything, so some choices must be made.

To some degree, it is quite enough to know how to use a tool to get the work done, rather than to understand all the details of how the tool works.  You don’t need to be able to build a computer to program a computer, and you don’t need to know all the tricks needed for sequence alignment to align sequences.

You just need to know enough to be able to use the tools: know 1) when it is appropriate to use a given tool, 2) how to run it, and 3) how to recognize if the results are sensible.  If you use the tool as a black box and don’t know what is going on under the hood, 1) and 3) is really important.

There are assumptions about the data underlying any tool, and if you don’t know what they are, you shouldn’t use the tool.  For example, if you do a statistical test on data that doesn’t look at all like the test expects, it is garbage in – garbage out.  Don’t do ANOVA on data that doesn’t look normal distributed.  You will get significant results all over the place, but they will be artifacts.

What should we teach, then?

I’m not really sure, but for mathematics I think some basic statistics is really essential.  Not all the arithmetic involved, that isn’t really that important if you have tools for doing it anyway, but the assumptions underlying some basic tests, the importance of checking these assumptions and what significance really means.  This is essential for any data analysis, and everyone needs to know it.

For computer science, probably a bit of complexity theory.  Enough to know that some problems are feasible to solve and some are not, and have some intuition about which kind of problems fall in which category.  Probably not much more than that for complexity theory.

Some programming – script programming – just to make it easier to manipulate data.  Manual manipulation of data is just too tedious and error prone.

If you expect ever to have to implement some analysis yourself, then also some basic data structures and algorithms.  Just the very basics.  You need to know much more computer science to build really efficient tools – you probably need to come up with some new algorithms and some algorithmic engineering – and if it is not your main field, then you are probably better off getting a computer scientist or engineer to look at it.

Understanding the math is probably a lot more important than understanding the computer science, for a biologist.  At least one not specialising in computational biology or bioinformatics.

–

213-210=+3

I got an email this morning pointing me to Scitable at Nature.

What Is Scitable?

A free science library and personal learning tool brought to you by Nature Publishing Group, the world’s leading publisher of science.

Scitable currently concentrates ongenetics, the study of evolution, variation, and the rich complexity of living organisms. As you cultivate your understanding of modern genetics on Scitable, you will explore not only what we know about genetics and the ways it impacts our society, but also the data and evidence that supports our knowledge.

I browsed around a bit and there are a nice collection of short articles describing various areas of genetics and genomics.  I haven’t tried out any of the learning paths or any of the other features yet.  I have to finish a grant proposal so that has to wait.  Anyway, you might find it interesting to check out.

–

195-199=-4

I’m reading on teaching right now, going through links provided by Peter Beattie here.  Quite instructive.

Anyway, that is not what this post is about.  I just saw this one quote from the piece I’m reading right now and wanted to share it.

In fact, students, too, may resist authentic teaching — at least at first. For one thing, they may prefer to avoid unnecessary intellectual challenges such as those entailed by a more active, probing form of learning. The introduction of a nontraditional science program led one 10th grader to exclaim, “We see what all this is about now. You are trying to get us to think and learn for ourselves.” Exactly right, replied the teacher, relieved and grateful that the message was getting through. “Well,” the student continued, “we don’t want to do that.”

Indeed.

I hope that is the problem I am having lately.  Otherwise I’m just a crappy teacher – something I cannot rule out – which would be pretty bad for my chosen career.

–

188-196=-8