Re: [xml-dev] Are you an XML scientist? Do you observe data in the wild

[Norman Gray <norman@astro.gla.ac.uk>]

Roger, hello.

On 2014 Apr 27, at 10:40, "Costello, Roger L." <costello@mitre.org> wrote:

> A physicist observes nature and then creates a model of what he observed. The model often takes the form of a mathematical equation.

Well, there are a couple more steps in there, perhaps.

Realise X is interesting.
Mathematically model it in the simplest way plausible.
Calculate.
Write a grant application.
Calculate.
Four weeks later predict that the entire universe is a blob of radioactive green goo.
Which is an over-simplification.
...or at least not terribly interesting.
Write another grant application.
Add another feature to the model.
Four weeks later decide that this is now too hard to solve.
Four weeks later than that, decide that no, it really _is_ too hard to solve.
Have a drink.
No, come on: it's almost there...
Spend two months writing a program to automatically reduce hypergeometric functions to their simplest form
Triumphantly plug that back in to the calculation and discover, only two weeks later
...the entire universe is a blog of radioactive _red_ goo.
Which still isn't _quite_ right.
Oh well, the reduction algorithm was interesting, so see if you can publish that.
Decide to solve it with a simulation.
Write another grant application.
Three years later discover that you have a monster Fortran code which solves Y...
...and seems to show that X is impossible.
Oh well, that's still a couple of papers.
Decide that Nature is telling lies.
Write another grant application.
Ah, OK, the experimentalists have just measured the wrong thing.
...or you misunderstood what they said because *mumble*.
Your code now matches what the experimentalists now say Nature does,
...Success!
But because you've just got here by simulation, you basically still don't understand what's going on.
Have another drink.
Discover that a year ago, while you were still head-down in the code, X was shown to be a non-issue.
Which, you decide, you'd predicted all along, so you write that up, too.
Write another grant application.
Discover you are now an expert in the field of monster codes solving Y.
(was that the plan? erm...)
Decide that this is a vital new research frontier and try to work out how to fund it.
Write another grant application.
Have another drink.
{da capo}

Your milage may vary.  Contents may settle during transport.  Do not calculate while operating heavy machinery.  Do not give children unsupervised access to Fortran.  Nobel prizes not to be taken orally.

> Here's another relevant passage from the book that I am reading:
> 
> 	Scenario: you observe a pattern in the 
> 	strings that you are dealing with. So you 
> 	create a grammar to describe the pattern.

Woah -- computing is easy!  Me likes!

When it comes down to it, I believe that more thoughtful computing, like science or poetry or most higher-level intellectual attempts, is a muddled mixture of creativity, pig-headedness, hunches, careful ignoring of reality at the right time, careful accepting of reality at a different right time, keeping your eye on the goal, adroitly changing the goal if you discover you have a really great solution to fun problem Y and can get away with it, getting funding, eventually retiring to take up beekeeping.

All the best,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK