My second Mac

Al Kossow aek at
Mon Jan 8 12:37:10 CST 2007

 >> Beyond that, you can easily spend more upgrading than a faster newer
 >> machine would cost you on the used market.   That said...
 > Nah, I'm just trying to figure out the Mac "culture".  I want to get
 > my currently unused workstation monitor on the G3 and poke around a
 > bit.

Forget about OS X then, and think about the evolution of a computer from
a 128K 68000 through a G4 with a gigabyte of memory.

There are still applications that will run on the last version of OS 9 
that were written for the original Mac.

Mac users would typically have many times the number of applications 
than a Windows user, mostly because most had no installation 
requirements other than dragging the app onto the system. Drivers and 
system extensions were easily added by putting them in the Extensions
folder inside of System. The way some of them worked (patching into
system calls) sometimes resulted in system conflicts.

The biggest problem in the 68K Mac world was a holdover from the 
earliest days of trying to fit into a tiny memory footprint. The
flat 68K address space was segmented to create position independent
code chunks that could move around as the system compacted memory.
Unfortunately, the data space was also segmented, so the APIs often had
64K restrictions on data areas.

PPC Macs got rid of that, and went to a flat memory model, which made
programming applications with large code/data footprints MUCH easier.

The big problem with the MacOS was there really was no architecture.

Features were tacked on, including things like two different shared
library architecutres on PPC because they evolved through two different
devlopment groups. In the end the API was a "Mactintosh Mystery House",
with dead-end APIs, and APIs that were bloated and ridiculously 
over-engineered (Comm Toolbox, for instance).

Adherence to backwards compatibility really was why it was impossible 
for Apple to come up with a replacement for MacOS until someone (Jobs)
finally cut the cord. This was impossible from the bottom up, no one at
the top was willing to have old applications break en mass.

More information about the cctalk mailing list