Microsoft’s April 1998 Meeting With Steve Jobs
Internal Tech Emails has an interesting letter from Microsoft’s Ben Waldman that discusses negotiations over ClarisWorks (Apple agreed not to advertise their own product and instead preload a video promoting Microsoft Office), QuickTime, and Internet Explorer (bundling and promoting it in exchange for Microsoft having it ready when Mac OS X shipped).
The meeting took place the month before WWDC, where Apple would announce that Rhapsody was becoming Mac OS X and that Carbon would be added. Waldman writes:
Of course, this won’t be positioned as Rhapsody cancellation -- they’ll say that you’ll still have OpenStep/Yellow Box, and be able to run it on Windows, and on MacOS, except that will be one year later, and that Apple will do the “right thing” and preserve peoples’ investment in MacOS, while still providing an advanced UI runtime for people who want it (and it will still be accessible from Java). While Rhapsody required developers to do a lot of work to get pre-emption and protection, in this scenario, Apple does “95% of the work.” Later, however, in a smaller group, I asked Steve point blank if he’d ever believed in Rhapsody, and he said “no,” adding something about his duty being to NeXT shareholders.
The way I’ve heard the story told before, Carbon wasn’t part of the plan until Scott Forstall argued for it after key developers rebelled at the idea of rewriting in Cocoa. But perhaps something like that was always at the back of Jobs’ mind. It’s not clear to me when work on it started.
At the time, Apple apparently hoped to get Mac OS X running on PowerPC 604 Macs with 32 MB of RAM, but it officially shipped requiring at least a G3 and 128 MB of RAM.
Previously:
Update (2021-07-15): See also: Hacker News.
2 Comments RSS · Twitter
>At the time, Apple apparently hoped to get Mac OS X running on PowerPC 604 Macs with 32 MB of RAM, but it officially shipped requiring at least a G3 and 128 MB of RAM.
Which is interesting. Those first specs are fairly close to the ones for NeXTSTEP 3.3. I wonder what changed to make system requirements so dramatically higher.
Aqua played a role, I'm sure. But was DisplayPDF that much more resource-intensive than DisplayPostScript? Did the Carbon layer add a lot of overhead? Were Mac users less tolerant of low performance than NeXT users / was NeXT always rather sluggish?
And it's not like (and it's not like 10.0 was… particularly fun to use on those latter specs. Even with the much faster 10.2, on a G3 with 384 MB RAM, it still didn't feel exactly zippy.
@Sören My recollection is that at the time the NeXT users couldn’t believe how much Apple had slowed it down. Aqua required more RAM and computation (which was not yet hardware accelerated), and Foundation was rewritten on top of the new Core Foundation, with the extra layer reducing performance.
