12 Oct
2008
12 Oct
'08
7:08 p.m.
On 12 Oct 2008 at 19:38, Richard wrote:
...and I find batch mode to be highly frustrating. I
learned
computing in an interactive environment on terminals and I never want
to go to a batch environment, I always prefer the immediate feedback
of interaction. Sometimes even waiting for a compiled language pisses
me off :-).
Another case of WYKIWYL - what you know is what you like. Even
today, with all of the fancy development environments, I develop
essentially in batch mode. Lots of desk and paper-and-pencil work
before writing the first line of code. When a snake shows up, go
back to the desk and run the code through in your head until things
make sense.
I find using an interactive IDE to be almost counterproductive--I
can't keep my thoughts as well organized.
When Sunnyvale was doing OS work for the CDC STAR-100, we used a
couple of hardware emulators, STAR-1Bs, that ran at perhaps 1/100-
1/500 the speed of the real thing. Used the same peripherals and
stations, just a great big pile of microcode to do all of the CPU
stuff. (CDC eventually disposed of the things with two CEs working
for a week with bolt-cutters and sledge hammers reducing them both to
scrap. I still have a rectifier and heatsink from the PSU).
Making an OS change and recompiling the kernel was a "submit your
changes before you leave at night and hope the all-night compile
works--and the machine stays up". Boot your new kernel, watch the
system crash and burn and figure out why just in time for you to
submit your changes for the next all-night compile.
The only alternative was to hope you could snag some time from one of
the STARs at LLL. Or catch a plane to Arden Hills and use the STAR at
ADL.
You learned to be very careful--and very patient.
Cheers,
Chuck