30 August 2007

On drowning fishes ... (1)

Any blogger with a scientific bent, looking for a place of entry, can find an embarrassment of riches in any post by Dr C. His latest, Data, data, everywhere, enough to drown a fish..., is no exception; any reasonable person could spend a lifetime following the openings which it presents. Alas, I have only a few minutes and so will have to be selective.

One issue which Dr C raises, en passant, is the question of relative benefits between computerised and hands on data collection. This is something about which I feel strongly; I'll come back to it (later today, with luck) as it requires more time than I have at this moment and access to material stored on a server not immediately available to me. For now, I'll content myself with a small note on a sideways issue: Dr C comments that "I didn't see it [STELLA] in the list of programs discussed by the Growlery probably because it is a different field."

It's true that STELLA (a modelling software environment from ISEE Systems) is an excellent tool for getting to grips with complex data analysis. It's also true that it wasn't included in the article which Dr C mentions (Getting to grips with power tools) for several reasons. Modelling software is, as Dr C suggests, slightly different from (though not at all unrelated to) analysis.

In the case described, a model has been constructed within which the user can perform comparative analysis within an automated model. This does require either a preëxisting comfort level with the idea of software models in general or a preconstructed model with which to work. And, of course, you also have to understand the idea of comparing reaction curves to start with. (As an aside, one of the loveliest descriptions I have encountered for the calculus approach to modelling is given by Lakshmi, a fourteen year old school student, on the SCW Education site: "Calculus is when you imagine very small bits of a problem so you can get your head round it, then imagine that small bit happening over and over again, forever, to make it back into the big problem again but now you understand it.")

There are several environments of a similar type to STELLA, all of them good in different ways, and it pays (if you have the opportunity) to look around for the one which suits you. These days, UML (Universal Modelling Language) has brought them all into closer coöperation with each other. Anyone interested in my take on this area may like to read Why is Lego the best toy in the world?, written last year.

No comments: