Re: Randy's inherent hate of Apple IIs

[toddpw@ugcs.caltech.edu (Todd P. Whitesel)]

shack@onyx.southwind.net (Randy Shackelford) writes:

>This would be a good reason, aside from one detail: system 4.0 the first
>release of GS/OS, came out in '88. System 6.0.1, the last release, came out
>in '93. That gave the engineers almost five years to fix any deficiencies.
>If you ever used 4.0, you know what a vast improvement in speed came with
>5.0. It seems far fetched to me that while they were making faster toolsets,
>finder, etc, they purposely left this supposedly ill designed FST mechanism
>in place.

I was an intern at Apple during the summers of 89 and 90. They had a couple
dozen people working on HyperCard GS and support projects for it (better
code generation by MPW IIgs Pascal for example). They had maybe five or
so people working on the O/S. These people spent most of their time getting
specific functionality in place, to support what was implicitly the
end-of-life phaseout of the Apple II; I remember seeing many people visibly
uncomfortable at large company gatherings because management continually
refused to make clear statements about policy with respect to the Apple II.

It is my opinion that many throughout Apple knew exactly how the Apple II
users felt yet the compelling business logic was to drop the Apple II as
quickly as possible without losing customer loyalty. But Apple II users'
pride in their machine made that extremely difficult. The resulting conflict
of interest and its eternal quest for resolution produced the incredible
waffling behaviors that we remember bitterly now.

If anything the FST interface was a means to an end -- good disk interchange
for the filesystems Apple cared about. It was not and was never intended to
be a contribution to computing history. As such there was no incentive to
do anything more than what was required to get it working and shippable.

Sorry Nathan, Apple didn't "screw up" the FST interface. They simply never
finished it. Sure there was talk of doing it (what do you think I was asking
people while I was there?) but it always went something like this:

    Why don't we publish the FST interface? It's really cool.

    It's just not worth it. It means committing people to support it,
    fix bugs, add features no one knows we'll need until after the golden
    master ships to 3rd parties and the little guys really start messing
    with it... It's just not worth it. No one knows how much longer we'll
    be maintaining GS/OS anyway. And there are so many other things we can
    be working on that we know lots of users will enjoy, like the DMA SCSI
    driver structure -- heck, 3rd parties will be all over that.

    What's so hard about doing a good job but then releasing it "as is" ?
    Hey, we're Apple. We know what we're doing. Right?

    Yeah, but that's never good enough for 3rd parties. If we do everything
    for them, they scream we didn't give them any opportunities, but if we
    don't spec literally everything they hammer us for doing a shoddy spec.
    AND everything we spec is frozen in stone. Do you know how many weird
    things had to be generalized when we put the HFS FST in? Just dealing
    with the colons in the pathnames required corny hacks so we could be
    backwards compatibile with prodos 16. And there are still ambiguous
    pathnames because of the numbered prefix scheme -- those were invented
    before anybody thought of the FST idea in the first place. You know, 3rd
    parties still blame us for that sort of thing as if we're supposed to be
    omniscient or something. At least on the Mac we got everybody trained to
    avoid that sort of mistake early on, and it's only a threat when you use
    the MPW shell.

    Okay, but there are some significant products that could really benefit
    from it. There's this GNO thing out there that lets you do a lot of unix
    stuff on top of GS/OS, but they really could use things like links and
    owner/permission bits, even if it was hacked with the FSTSpecific calls.
    If we just published a spec they could write to, they really wouldn't
    have much to complain about, since we gave them all this power.

    Something they could write to? Do you know what you're saying here? How
    are you going to make sure GS/OS doesn't have any latent bugs that get
    uncovered by adding links, until after somebody goes out and does it and
    a lot of users beat on it? It's not just links. Any new concept we add to
    the interface might do it. Suppose somebody wrote an FST that operated on
    shrinkit archives -- now you're re-entering the FST dispatcher and of
    course that won't work but "we're Apple we do everything the first time
    right?" give it up, this is the real world. After HyperCard GS is done,
    the only committed people for Apple II work is probably going to be a
    small maintenance crew. Just forget about it and focus on something that
    doesn't have so many things depending on its stability, okay?

It can go on indefinitely, but I think this says enough.

Todd Whitesel
toddpw @ ugcs.caltech.edu