Me: Computers are getting faster

Daughter: Ours?

Me: No, new computers

Daughter: Oh…

The truth of the matter is that we can spend all the time we want worrying about the ‘coming software crisis’ but really the vast majority of computers are not getting any faster. The frustration of the average user is clearly evident when they complain about the speed of their machine and that they just can’t run the current software. What’s amazing is how fast these average computers really are, so fast that it’s hard for us to really get our heads around that speed at the hardware level (can you really visualize a CPU doing 400 MFLOPS at the instruction level?).

As an industry we need to be much more conscious of our design decisions. We should concentrate on listening to our customers rather than relying on the prevailing wisdom of Moore’s Law to help us along. Our customers don’t care about .NET or latest-greatest-programming-language-x or multi-core processors, they just want the programs to be stable, fast and functional.

There’s a humorous story floating around Microsoft that Bill Gates would occasionally show up to ridicule some bloated feature or another saying something along the lines of, “You’re an idiot, I fit a full BASIC implementation into 4K“. Maybe it’s a myth from the early Microsoft days but I always liked that story.

http://www.interact-sw.co.uk/altair/index2.html

“SyntaxCheck (RST 1)

Here is a truly beautiful piece of code, it’s Golden Weasel richly deserved. It’s used at run-time to check syntax in a very cool way : the byte immediately following an RST 1 instruction is not the following instruction, but the keyword or operator ID that’s expected to appear in the program at that point. If the keyword or operator is not present, then it Syntax Errors out, but if it is present then the return address is fixed-up – ie advanced one byte – and the function falls into NextChar so the caller has even less work to do. I honestly doubt syntax checks could be done more efficiently than this. Sheer bloody genius.”

The introduction to the usage of 8080 Restart functions is also particularly interesting.

“After the TMG version of B was working, Thompson rewrote B in itself (a bootstrapping step). During development, he continually struggled against memory limitations: each language addition inflated the compiler so it could barely fit, but each rewrite taking advantage of the feature reduced its size.”

…

“C is quirky, flawed, and an enormous success. While accidents of history surely helped, it evidently satisfied a need for a system implementation language efficient enough to displace assembly language, yet sufficiently abstract and fluent to describe algorithms and interactions in a wide variety of environments.”