Larry McVoy lm at mcvoy.com
Tue Sep 19 03:40:11 AEST 2017

On Mon, Sep 18, 2017 at 01:25:41PM -0400, Clem Cole wrote:
> On Mon, Sep 18, 2017 at 12:46 PM, Steve Johnson <scj at yaccman.com> wrote:
> >
> >
> > ???...
> >  After the first time, there is little to learn, and the tedium of
> > debugging a compiler and OS on a bare machine, when the documentation of
> > the machine was hastily written and often incomplete, was frustrating
> > almost beyond describing.
> >
> ???+1??? And the HW does quite work the way it claimed too....
> It's fun ... once  :-)

Wussies, I did it twice :)  Not really true, I worked a port to the ETA-10
and managed a group that was doing a bringup on a MIPS chip (Cobalt Qube).
I had more of a clue the second time around, actually quit over it.

The details, if you care, were that I was brought in as the director
of software, there was no VP, I reported to the CEO.  I was assured 
that I would be treated like a founder, yada yada.

OK, some time goes by and the CEO wants a schedule.  Be aware that
things were a mess, no backups, no nightly builds, no source control,
engineers were hoarding "their" code on their machines.  It was little
disfunctional, but I was straightening all that out, had it pretty
much under control by the time they asked for a schedule.

We were using GCC which hadn't had a ton of MIPS time under its belt,
and Linux, which was in the same state.  We could boot a kernel but
it crashed (I think, this was a long time ago).

Anyway, they wanted a fast schedule (don't they always?) and I looked
around and said "6 months".  Which was aggressive but I felt it was
doable, the kernel was booting, once you get that stable, userland tends
to fall into place.  We were also developing a web interface to the box,
that was the whole point, it was a "web server in a box" with a web UI,
no login for you (well there was but we weren't supposed to use it).

They freaked at the schedule.  "The board will never go for that!"

So I called Bill Earl, who has done more bringups on MIPS chips than
anyone, period.  At least so far as I knew at the time, still don't
know of anyone who has done more.  He asked a bunch thoughtful questions
and said "weeeeell, maybe you could do it in 3 months.  If you don't hit
any bugs."  I asked "Would you bet your kids college education on it?"
He said "Oh, hell no, I wouldn't bet $20 that you'll make it in 3 months.
You hit one compiler bug, that could blow the whole schedule easy."

So I go back to the CEO and said "6 months".  He's all pissed and says
he'll take it to the board.  I say I want to be there to make my case.
He says "Nope, you aren't an executive."  I handed in my resignation.
Founder my ass.


More information about the TUHS mailing list