Comments on: The Second Law of Software Design

By: Max Kanat-Alexander

Max Kanat-Alexander — Mon, 10 Mar 2008 11:15:14 +0000

Thanks! Well, if a philosophy does describe the fundamental nature of the universe, it ought to be useful everywhere, eh?

-Max

By: Claudio Criscione

Claudio Criscione — Mon, 10 Mar 2008 08:17:13 +0000

Interesting ideas behind these laws, and it is somewhat funny to see Panta Rei, a statement some thousands years old, being used in such a context No joke intended, I find it really nice to see that we are using Greek philosophy in software engineering!

By: Max Kanat-Alexander

Max Kanat-Alexander — Fri, 07 Mar 2008 19:52:00 +0000

Ah, okay. Yeah, an understanding of the concept of a symbol is necessary to an understanding of computer science, although it would be somewhat irrelevant in software engineering/design, except that it’s a total requirement to understand any programming language.

-Max

By: Alex Vincent

Alex Vincent — Fri, 07 Mar 2008 19:48:19 +0000

I was thinking along the lines of how object-oriented programming, or even procedural programming, involves the use of variables - and if you don’t understand that a variable can have its value change, then you’re not going to get very far in programming.

It goes beyond that, though, in the concept of having letters (or in modern languages, whole words and pseudo-words) represent values and objects.

By: Max Kanat-Alexander

Max Kanat-Alexander — Tue, 04 Mar 2008 06:04:21 +0000

Hm, algebra in what sense?

-Max

By: Alex Vincent

Alex Vincent — Tue, 04 Mar 2008 05:28:58 +0000

I look forward to reading these laws, and wonder if algebra will make an appearance among them on a list of software design prerequisites.

By: Max Kanat-Alexander

Max Kanat-Alexander — Mon, 03 Mar 2008 22:36:21 +0000

Yes, you possibly could! At least, up to a certain point. I mean, there are some things about the future that you just can’t predict or design for. For example, before there was Unicode, it would have been impossible to design your program to deal with Unicode, so there’s still going to be changes like that–unknown and unexpected things that you have to allow for. Instead of trying to predict the future, I usually just try to make my software easy to change, so that when it is the future, it’s easy to accommodate the exact demands placed upon the system.

-Max

By: Max Kanat-Alexander

Max Kanat-Alexander — Mon, 03 Mar 2008 22:34:18 +0000

Yes, I know the law of entropy, and I could have used it, but one of my goals with these articles is to be simple, to use ideas and principles that anybody could understand, without having to teach them specific laws of physics or anything like that.

Anyhow, I don’t think I incorrectly involved physics. My first statement is equivalent to the statement, “It is impossible for matter to reach absolute zero.”

-Max

By: Riyaz

Riyaz — Mon, 03 Mar 2008 19:45:03 +0000

Given that it takes more time to change a piece of software compared to the time it takes for the environment around it to change… can we draw a corollary that “A software is said to be well designed if it survives environment changes without having to change itself”?

By: LpSolit

LpSolit — Mon, 03 Mar 2008 19:13:29 +0000

Rather than incorrectly involving physics here, you should simplify your “demonstration” by invoking the 2nd law of Entropy, which would be much more rigorous: http://en.wikipedia.org/wiki/Entropy#The_second_law