Comments on: Computational science approach to programming

By: It gets compiled anyway | Polymath Programmer

It gets compiled anyway | Polymath Programmer — Wed, 02 Jul 2008 09:02:32 +0000

[...] feel I misrepresented my intentions in a previous article. In that article, I gave a code [...]

By: Vincent Tan

Vincent Tan — Sun, 29 Jun 2008 13:18:08 +0000

For bigger arrays, by all means go ahead with iterating variables. It’d be a more effective method than changing numerical indices.

I just feel the example and it’s associated context (smallish array) works fine with hard-coded numerical indices.

By: Ben Barden

Ben Barden — Sun, 29 Jun 2008 12:23:40 +0000

I might use the c[i++] method for a bigger array, as adding items in the middle could get annoying as I’d have to keep changing the indices… Then again, I could sort the array… or do something more like this: array(‘a’, ‘b’, ‘c’).

By: Vincent Tan

Vincent Tan — Mon, 23 Jun 2008 13:08:24 +0000

Christopher – I’m sure the character array example can be optimised by compilers. The thing is, I’m not that good. I assume the programmer had a good reason for writing it in that particular way. After I figured what’s going on, and discover that there wasn’t really any good reason, I get a little frustrated because of the lost time.

I’ve heard of F#. Just haven’t gone to explore it much. And yes, I’m a .NET programmer, dealing mostly with ASP.NET, though I do Windows and console applications too. That array example was paraphrased from a C program on Unix.

Aaron – Hey! I just read that article on your blog. Yeah it’s freaky about the coincidence…

By: Aaron Falloon

Aaron Falloon — Mon, 23 Jun 2008 11:50:58 +0000

Jeepers, we must think very alike because I posted a similar article a few days ago. I think it’s always helpful to define the problem and write that definition down. That way, you’re going to solve the actual problem.

By: Christopher Tay

Christopher Tay — Mon, 23 Jun 2008 11:43:38 +0000

I agree with you that a fatal mistake often made is to plunge straight in to coding without stepping back to reflect upon the proper algorithms or data structures to use. Having proper algos and data structures from the macro point of view have a much higher impact than the tiny optimizations to be made at the micro level. As to the character array example, I think most existing mature compilers should be able to reasonably optimize the code. But of course, the degree and forms of optimization varies amongst the different compilers.

If you’re more interested in coding to solve problems instead of the technical details associated with writing a correctly executing program, you might want to take a look at functional languages. I guess you’re doing lots of .NET? There’s the programming language F# which runs on .NET and is similar to ML.