Red Hat Diaries/0039

Definitive Technologies

> I have been reading your Blog on X and am finding it amusing and
> informative. That you are willing to start learning a new OS and
> development environment instead of defending what you know as the
> definitive technology (which is the most common adaptive mode
> amongst technologists) is commendable.

I don't know how far back things are down under, but Microsoft is 
certainly no definitive technology. Not today, not ever, and less so 
today than ever. What's a definitive technology? How does one 
describe a definitive technology? I think you need two comparative 
technologies to even whisper the term.

In the good old days IBM ruled the world. Everything was IBM 
mainframes. Our mainframe had over 3500 3270 terminals, and 3270s 
are not smart. They have a few KB to manage screen stuff, that's it. 
The CPUs were big enough to pitch a tent in. We had two, one was 
always backup, and if either failed IBM flew in another one from 
Paris within the hour. The whole computer covered a complete city 
block - and a very big city block at that. There was only one 
technology back then, and it was definitely definitive.

IBM may have been definitive, and I have a lot of respect for IBM, 
but it was not fun. Maybe if I had been in on the design from the 
beginning, but I think not. The architecture was an unending 
succession of bad design decisions which were necessary because of 
other older bad design decisions etc. No fun.

Along comes Unix and you have something to compare with. Brian 
Kernighan is perhaps the greatest teacher in the world, and the work 
place known as the CSRC at Bell Labs must have been brilliant. BASIC 
is gone, assembler is basically gone, we have something called C and 
it's brilliant. C lets you see how things work. It doesn't abstract, 
it doesn't numb, it doesn't make you clueless. It requires you to do 
hard work, but who cares? C opens a world never before seen by many 
people.

When I had Unix and IBM MVS side by side, it was not hard to choose. 
But then again, when I had that old terminal based Unix and the PC 
side by side, it was not hard to choose either. Doing good 
non-trivial work on one of those old Unix terminals was almost 
impossible. There were surely secrets to it, but they weren't 
exactly being broadcast around, and the results were abysmal at 
best. The PC was a piece of hardware with a hardware interface. It 
didn't have an operating system. It was anarchy, and it was cool in 
that regard.

While we worked on the PC, things gradually improved on Unix, 
although we didn't see them. No one would ever claim they were not 
working on Unix, for even on the PC you were continually trying to 
bring Unix - a way of thinking more than an OS - into play. It was 
all Unix, and it still is, to this day.

Window systems never offered anything ordinary systems could not 
offer. The kind of file management you can get with a window system 
is available with ordinary systems. While some of us were looking 
suspiciously at Presentation Manager, others were commenting that 
none of this was beyond the reach of single task systems - except 
the multitasking stuff, which at best is simply a way for users to 
confuse themselves, have a lot of windows cluttering their desktops, 
and drool over a mouse. Programs like Lotus 123 and SideKick paved 
the way to multitasking, and yet multitasking as we know it today is 
primarily The Day of the Locusts: The clueless of the world have 
attacked, and we're going down for the count. Even SideKick and 
Lotus 123 handled data transfers in single task systems. The basic 8 
bit character set was enough to decorate any window, any popup, 
create any 'menu'. We haven't really seen anything new.

Work at PARC gave us the mouse, the window - and SmallTalk. 
SmallTalk resulted in Objective-C. Bjarne Stroustrup somehow ended 
up at Cambridge, and after that at Bell Labs, and when he'd flopped 
with his first program, he set about creating a new language which 
would make his program run faster. Bjarne lives in Simula, which is 
not the same thing as SmallTalk, and again you have comparative 
technologies.

Apple defined the graphical user interface for the public with the 
Macintosh. Apple took what were research products such as the mouse 
and bit blitting and made them into commercial products, geared to 
the consumer both in terms of usability and in terms of price. 
Presentation Manager, VisiOn, GEM, Windows and all of them tried to 
follow suit.

Jobs had Sculley on board and Sculley ousted Jobs from a position of 
influence in his own company, so Jobs quit. And he took his vision 
with him. He founded NeXT Computer, and again you had comparative 
technologies. Already back in 1987 the system Jobs' new company came 
out with was better than anything available even today. NeXT as 
compared to anything since - Windows in particular, but also MIT's X 
- is infinitely superior, and represents itself a 'definitive 
technology'.

As does Objective-C. Which builds on SmallTalk. Which in turn is the 
definitive look into window systems. What multitasking we have today 
is the inheritance from the 'real' operating systems such as Unix, 
VMS et al. and has nothing to do with any windows. The two seem 
intertwined and may in fact be so, but their evolution is completely 
separate.

Once Adobe releases Photoshop for OS X it's all over. Once this 
program is up and running properly it's 'lights out' for all the 
rest. The GIMP may be good, but the pros want Photoshop. And as soon 
as the new Photoshop becomes the industry standard all these pros 
will upgrade to OS X. As once the Macintosh was a new product and 
not an established winner, so it is with OS X of today. But OS X 
will always be more. For it is NeXTSTEP to the core. Its entire 
class hierarchy is the NeXTSTEP hierarchy. Nothing has changed. 
Where competing technologies such as the MFC trip up and fail, 
NeXTSTEP, using Objective-C and based on the original SmallTalk idea 
of MVC - Model View Control - will continue to lead. And as if that 
is not enough, NeXT - that is to say OS X - has its old leaders on 
board at Apple today, and the old PARC SmallTalk people are there 
too. And today's OS X interface is 'show off ware', because it is 
not absolutely necessary but is saying, 'we can do this, and we know 
you cannot, eat dirt'. And it's true - no competing comparative 
technology can do what OS X can do today - what NeXTSTEP did already 
fifteen years ago.

If you accept the GUI as an interface standard; if you accept 
multitasking systems on home and office desk computers; then you 
have only one possible 'definitive technology'. And it does not come 
from MIT, it does not come from Bell Labs (even though Plan 9 could 
out-do them all with a prettier interface), and it certainly will 
never come from Redmond (heaven forbid).

It comes from One Infinite Loop, Cupertino.

Click here »