Welcome to the first in a hopefully long and entertaining series
with some of the wizards of the Macintosh programming world. Although
the Mac may be the world's easiest computer for the rest of us, it has
not traditionally been the easiest system to develop on. We all owe a
debt to those pioneering programmers who used their time and skills to
provide us with the Mac OS and so much wonderful software to run on
it.
Today's interview is with Tom Pittman. Tom has a long history in the
microcomputer industry. He is the co-author of a well respected text on
compiler design called, appropriately enough, The Art of Compiler
Design. He has taught Computer Science at Kansas State and
Southwest Baptist University. In the Mac world, he is best known as the
author of "CompileIt", an amazing program that allowed compilation of
HyperCard scripts into super-fast machine code.
Tom's website has his bio as well as many interesting links:
Comment from Tom: "I think the Mac is great to develop on;
it's just hard to develop for. What I call the 'Conservation of
Complexity': If a program is easy for the user, it's hard for the
programmer, and vice-versa."
Rick Lawson: Tom, thanks for agreeing to this interview.
Would you please tell us how you became interested in programming?
Tom Pittman: In high school, I wanted my own computer. I
collected electronic parts - 200 triode vacuum tubes - to build a
computer. I didn't have a clue. I was a math major in college - that
was before they had such a thing as "computer science" - but all the
proofs were getting to be a drag.
The math major included one lab course in programming. I loved it.
That summer I went to interview for a tech writer job in a local
government laboratory; they had already hired somebody else, but they
had an opening for a "mathematician". I told them that was my major,
and they hired me. They really wanted a computer tech, but the pay was
better for a mathematician.
Rick: What specifically what drew you to the Mac?
Tom: In grad school I was learning about operating systems
and windowing environments. I even started to write my own windowing
OS. To do my thesis, I needed some special arrow characters not in a
daisywheel printer, but the Mac had just come out, and I knew it had
software fonts, so I could create my own.
My prof told me I should try it before spending any money and loaned
me his. He never got it back. It was everything I wanted in a computer.
I wrote my own native-code Pascal compiler. That was when everybody
else was buying a $10K Lisa to program
for the Mac. I was going to write programs for the Mac, but for two
years I didn't need to, because everything I wanted already existed on
it.
Rick: HyperCard was really a
groundbreaking program in opening up software development to the
masses. On the limited hardware of the time, HyperCard was criticized
as being too slow. CompileIt directly addressed this issue, as well as
providing access to low level Mac ROM routines from HyperCard.
Please tell us how you conceived of CompileIt and some of its
history?
Tom: My specialty is compilers. When HyperCard came out,
Bill Atkinson
said there would never be a HyperCard compiler. Strictly, he was right,
because you can write self-modifying scripts in HyperCard, and that
just can't be compiled in any reasonable way. But I figured that most
of what people do can be compiled, and I could punt the rest. Besides,
it would be fun to prove Atkinson wrong.
Pundits were also saying that HyperCard was too limited to do useful
things. I decided to prove them wrong at the same time by writing my
compiler entirely in HyperCard. I did that. I never used any other
programming tool besides ResEdit (and a text editor) for any version of
CompileIt, not even the first one, which I wrote completely in
HyperTalk. It was incredibly slow, but it worked. Then I compiled it in
itself, and it got much faster.
Around the time I got the idea of HyperCard compiling HyperCard,
Macworld magazine came up with their "SuperStacks" HyperCard
programming contest. CompileIt was barely functional when I submitted
it, and the philistines in San Francisco didn't give it the attention I
thought it deserved. By then I'd put enough effort into it that
somebody mentioned me to Brian Molyneaux at Heizer Software, and he
immediately saw its advantages and became my publisher. Brian was great
at seeing how it could be improved.
Rick: Who were some of the more interesting users of
CompileIt? How about unusual or interesting uses?
Tom: I really didn't have that much contact with the users;
Brian did all that. I answered questions on the CompuServe HyperCard
forum until their system software locked out text-only users like me. I
did once go to an Apple Developer's Conference, and over lunch one day
I happened to strike up a conversation with one of my users. I didn't
know who he was until much later, when I heard of The Hitchhiker's
Guide to the Galaxy. Douglas Adams was an
avid CompileIt user. I also took apart the Mac version of Myst (the
game); it was implemented in HyperCard, and I found one of their XCMDs
had been compiled by CompileIt (I can tell from the machine code what
language it was written in).
Rick: Did Apple ever approach you about integrating or
bundling CompileIt with HyperCard?
Tom: Apple never understood what HyperCard was. Rumor was
that they only offered it because Atkinson wanted people to use it, and
if Apple didn't agree to ship it free with their computers, he
would go somewhere else. They did not want somebody as smart as
Atkinson working for a competitor. As soon as their "free" agreement
expired, they started charging money for it, and they eventually
discontinued it.
Rick: You have some fairly strong opinions on OS X. Please
give us your thoughts.
Tom: I spent 20 years vigorously avoiding Unix - and failing often enough
to remind me why it was a good thing to avoid. If you must use
Unix, Apple's AU/X was
arguably the least painful way to do it (much easier to use than
OS X), but even AU/X was so bad it didn't last long.
After Apple announced their intention to replace their "aging"
17-year-old Mac OS by a "modern" 35-year-old Unix with a thick layer of
pancake makeup, I tried to get some version - any version - of
Unix/Linux up on
any platform. For over three years I failed utterly to get
anything to survive the first reboot. Even OS X crashed "kernel
panic" dead on me the first day.
When I got on faculty at the university, there were gurus there who
knew the Unix system well enough to get it up for me.
Unix is a system designed by geeks
for geeks; it is not a system for real people.
Unix is a system designed by geeks for geeks; it is
not a system for real people. The only people recommending OS X
are geeks whose paid job already involved using lots of Unix computers
before they ever looked at OS X.
Recall that I "are" a geek myself (PhD in computer science). But
like the rest of the real world, I want my computer to help me get my
job done, rather than forcing me to learn arcane commands that don't do
useful things. The Mac OS gave my productivity an order of magnitude
boost; everything since then (even Mac OS versions after 6) only goes
backwards.
Rick: Do you think programming is an art or a science?
Tom: Some of each. The painter must know his colors and how
to make perspective and light sources work; that's science, and he
can't really make credible pictures without it. But nobody buys
paintings by engineers. You have to be able to visualize a world that
doesn't exist and then create it. As some sculptor reportedly said,
"You take this block of marble and remove everything that isn't the
statue."
I can teach principles of software design, but most people just go
through the motions, following the recipe; they don't really get
it.
Rick: Now that the software industry has matured, do you feel
software can still be a force for change in the world ?
Tom: That question seems to personify software. It's a tool,
not a force. People use tools to change the world. Computers -
and the software that makes them work - seem pervasive today, but I
think we are about where cars were 60 or 75 years ago. There is more
change coming, caused by computers, but not so obviously.
Rick: What are you working on now?
Tom: A sort of super-compiler. I'm using compiler technology
in natural-language translation. I have it working, sort of, but not as
easy to use as HyperCard or CompileIt :-) See http://www.BibleTrans.com/ for
details. The first prototype was written in HyperTalk (and compiled in
CompileIt), but now I'm trying to get it up on a PC.
In my spare time I also think about finishing that windowing OS I
started to write in the early 1980s. I want my Mac back. OS X is
not a Mac.
Rick: Any parting thoughts?
Tom: In the spirit of HyperCard/CompileIt (and TinyBASIC before that), I
wonder if it might be possible to get a Mac-like (non-Unix) OS up,
perhaps building on the open-source Linux kernel but without all the
Unix mistakes that make Linux unusable.
Think what would happen to Vista and Linux if that $100 computer
they are working on for third-world users was as easy to use as the Mac
was. Would some foundation fund development? Once we had version 1.0
working, it would acquire a life of its own. It can be done!
Rick: Tom, thanks so much for your time, and good luck in
your future endeavors. 
About the Interviewer
Rick Lawson is a Senior Developer specializing in
Web technologies. He lives in South Carolina with his wonderful wife,
son, daughter, and three cats. He has been enamored with Macs since
coming across a brand new Mac Plus in
college almost 20 years ago.
