Re: [tlug] Speaking of computer usage ....

["SL Baur" <steve@example.com>]

On 2/29/08, Stephen J. Turnbull <stephen@example.com> wrote:
>  There are still a lot
>  of people who claim to be "computer professionals" who don't
>  understand that, whatever the advantages of commercial OSes may be,
>  reliability is not one.

That was the case 2 decades ago.  I took great pride when I was
prototyping user interface software on Sun 3s to put a "panic" button
on every screen.  People would ask me, "what happens if I hit the
panic button?" and I told them to go ahead and try it and of course,
Sun OS would panic and reboot.

I haven't had a Sun of any kind crash on me since my TRW days, 17+
years ago.  I used a Tru64 Alpha almost every day and had a Sun
box as my main workstation at ETL in Tsukuba (until it was replaced
with a Linux box) and they never crashed.

Commercial Unix software is quite good.  I think it comes with the
territory.  We have a clean architecture, well-understood security
issues and we've solved them.

When we try to cut corners (can anyone say vmsplice(2)?  I knew
you could), we run into problems, but the basic architecture is
fundamentally sound.

Under normal usage, a computer should never crash.  Never.

Microsoft has turned back the clock to a time when they used to
run a job 3 times in succession to see if it gave the same answer
(in those days, it usually did not ...).

The irony of all this is that Microsoft was headed our way at one
time.  PC DOS 2.0 introduced all kinds of Unixy system calls
versus the CPMish PC DOS 1.1 ones.  At the time Microsoft owned
Xenix and the general word was that MS DOS was headed towards
a merger with Xenix.  If they'd actually done that, they would have
wiped *everyone* out.  Unfortunately for them, they sold Xenix to
SCO, continued on a divergent path and the rest is history.

Even giants make mistakes.  NIH is folly.

-sb