The Art of Developing Software

The below is taken from an interview conducted at the Computer History Museum in 2004 by Grady Booch. He interviewed early Apple developers Bill Atkinson and Andy Hertzfeld about the development of the original Apple Mac, and in particular MacPaint software.

On the art of developing software

Grady Booch: So what makes software beautiful for you? Do you care about the beauty of software?

Bill Atkinson: Oh yes, it’s an art form, like any other art form, and it’s not just practical, as in, does it do the job? Is it clean inside, does it look--? I would spend time rewriting whole sections of code to make them more cleanly organized, more clear. 

I’m a firm believer that the best way to prevent bugs is to make it so that you can read through the code and understand exactly what it’s doing, that there’s nothing tricky in it, that it’s all very straightforward. And maybe that was a little bit counter to what I ran into when I first came to Apple. 

There were a lot of people who prided themselves in how this little piece does something that I can’t for the life of me figure out what it is, but it’s magic that it does it. I would do things like deliberately assign something into a variable, and instead of putting ten operations concatenated onto one line, I’d use intermediate variables so I could understand the purpose of each of these. And I liked it. 

I found that if I spent time going over the code, cleaning it up, making it sometimes tighter – I did a lot of measurement and tried to make the performance good – but also making it so that it was straightforward so another person could follow in my footsteps, then I would feel proud of it.

Grady Booch: Don Knuth was telling me that he felt that MacPaint was perhaps one of the most beautiful programs he has ever seen, just in terms of its readability. 

On the MacPaint development process

Bill Atkinson: I had a different method of software developing than Randy Wiggington who was developing MacWrite. We had two different philosophies. Mine was you don’t get to add any new features until what you’ve got is working reliably, solidly; and his was, let’s add all the new features and then let’s debug them. I just think that that’s a bad way to do it because finding bugs is a very unlikely proposition. If you search really hard you might find half the bugs that are there, and the others are going to bite you and you just don’t know when. So I had a little more conservative approach to design.

Grady Booch: What you describe is very similar to what the Agile community speaks of these days; always have something executable through incremental development. And it’s very cool.

Andy Hertzfeld: Yes, yes, that’s really how the Mac was developed.

Grady Booch: You were the first Agiles. So by my count, or others, MacPaint was about 5,804 lines of Pascal and so on; 2,738 lines of assembly language.

Bill Atkinson: Plus all of the QuickDraw I’d built on top, because without QuickDraw you couldn’t do all that stuff.

Grady Booch: And pretty much everything had built upon QuickDraw.

Andy Hertzfeld: Yes.

Bill Atkinson: There were 70,000 lines of assembly language [in QuickDraw]. 

More reading
Computer History Museum: Oral History of Andy Hertzfeld and Bill Atkinson
Revolution in The Valley: The Insanely Great Story of How the Mac Was Made, by Andy Hertzfeld
Folklore: Anecdotes about the development of Apple's original Macintosh
BYTE Magazine (1984): An Interview: The Macintosh Design Team

Follow @dodgy_coder on X