success (was: On Schopenhauer and Market-driven code)
Jecel Assumpcao Jr.
jecel at merlintec.com
Thu Mar 16 02:00:02 UTC 2000
I hate it when the grownups are distracted trying to find directions on a map
and the kids in the back seat start punching each other ;-)
Now that things have calmed down, I hope you all won't mind if I make some
comments on Lawson's point about success.
First a little story. At the 1986 "National Computer Fair" in Rio de Janeiro
(Brazil), I was showing a little Smalltalk based Network Computer (no, it
wasn't yet running Smalltalk at the time, just a graphics demo) in the "VIP
booth" of a local PC maker. Since they had announced an AT clone but it wasn't
finished on time, I spent most of my time saying "No, this is not the famous PC
AT" instead of explaining the virtues of object-oriented programming languages
or the advantages of the Ethernet. But what I remember most was a guy who asked
me if it would run Cobol. When I said no, he became very angry and gave me a
lecture on how evil what I was doing was. I didn't make any sense to me - all
around there were dozens of booths selling machines that ran Cobol just fine!
Why couldn't he choose one of those instead of wanting to change my product to
be like all the rest?
I thought the same thing about Macs and hardware slots. How many of the people
who complained so loudly about the lack of slots actually went out and got a
Mac once the NuBus came out?
All this just to make it clear that I understand the negative reaction people
get on this list when they ask to make Squeak more like this or like that. In
fact, I am very skeptical about converting people from Java, C, Python or
whatever by adding features to Squeak. But I don't worry about that - I am sure
that most people who will be programming in 2005 don't program today. If we can
somehow get to them first, it will require much less energy than converting
existing programmers. Even that will take major marketing, however.
But why bother, you ask. As long as we have a tool that does what we want it to
do, isn't that good enough? It is for me, but there is a major problem: if we
let other people succeed (in the sense of getting more users), we will waste
more and more of our time getting along with them than doing our real work!
Imagine a world in which most computers were running Smalltalk instead of Unix,
Windows or MacOS. How much time would waste thinking about "native widget" and
similar? Or about converting JPEG and GIF to ColorForms? Or MIDI to Smoke? Or
XML to objects? Or talking to Corba or faking RMI?
I am particularly worried about MPEG-4, Active VRML and similar efforts. I
would hate for Squeak to have to include a full Java implmentation just to be
able to import these things. I would much rather that Netscape and Microsoft
were forced to include a Squeak plug-in with new releases so that their clients
can see the nifty new content being deployed on the internet.
Squeak isn't ready for the masses. Most people that I got to try my Slackware
Linux CD-ROM back in 1996 have since eliminated it from their machines and
don't use Linux currently. But I bet they would if they had tried it today
instead. Getting millions of downloads right now might not be the best thing
for Squeak. Yet that must be part of our future plans, or else we will be
doomed to try to reverse engineer RealVideo formats and ICQ protocols forever...
Does this make any sense at all?
More information about the Squeak-dev