[Coco] Question for those familiar with rbf.mn

[Gene Heskett]

On Sunday 01 February 2015 11:37:41 L. Curtis Boyle did opine
And Gene did reply:
> Hmm… I didn’t think it failed, just returned the segment list full
> error - it may have left all the blocks allocated, instead of cleaning
> up after itself, though. We hit this a few times at work (probably the
> worst was when the Atlanta Braves sent a file that was about 10-11 MB
> zipped - it was huge fixed length text file, bigger than the size of
> our 80 MB hard drive at the time. I had to write a program to read a
> line at a time from the ZIP using pipes, strip out all the stuff we
> didn’t need, and then write out the shortened line, in order to
> extract the data we needed).
> 
> L. Curtis Boyle
> curtisboyle at sasktel.net

That sounds like an interesting time, working under a looming deadline 
pressure.  I think I did some of my best work doing that.

I can recall the time I rename copied the program I wrote to make an edisk 
for a Grass Valley Group video switcher out of a coco2, and using that 
code, modified it so that I could detect the error faster, and walked thru 
the switchers  digital data path much like traceroute until I could 
determine exactly which chip was the faulty one.  One chip in particular 
wore a grass part number, but I was able to call AMD, who were making 
memory stuff at the time, described what the chip did, and they then sold 
me a stick of them with their number on it, for about $3 a chip.  What it 
was, was a 4 bit wide bidirectional fifo, used 2 chips wide as a rubber 
buffer between channels. The grass version would forget to shift out 
frequently, a fairly high failure rate that in 10 years just about used up 
that stick.  But the AMD part Just Worked(TM).

Where grass thought they had a proprietary part, they gouged us 
unmercifully.  I bought the first one I needed from them but the C.O.D. 
was for $1650 when it came in the door.  I could spend a week on the phone 
looking for whoever made it for that sort of money and did.

Grass also house made some video speed op-amps on airmail stamp sized 
ceramic plates, which also had a high failure rate.  I called them, found 
they did have one on the shelf, it was not guaranteed to be good, and it 
would cost us $1500 to find out.   No returns if it was bad.  'scuse me? 

Went trolling in the op-amp section of the TI chip books I had, and came 
up with one about 100X faster, 1/20th the power consumption, single ended 
supply, and $1.13 in stick qty's.  Worked well except for one problem.

Because it was faster, the delay through it was less, about 10 degrees at 
3.58MHz, beyond the range of the channels delay trimmers to compensate.  
What I should have done but didn't consider at the time, was buy 4 25 pack 
boxes (TO-5 case) of them and replace them all which would have 
automatically brought it back in time with itself & all I would have had 
to do was stick about 30 more feet of cable in the subcarrier drive path 
to get it back in time with the rest of the house.  Hind sight, 20-05 of 
course. ;-)

But that also would have been at least a months work, working around its 
use schedule as there was only about 6 hours a day I could work on it, and 
4 of those were in the wee hours of the morning. At that time, I was all 
the engineering staff and other things would have interfered. But it did 
get me perms to go hire an assistant to take care of the PM stuff.

Yeah, those were my more "salad" days.  Retired for about 13 years now and 
enjoying the freedom to do as I please within budget.  Whats not to like?

[...]

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS