Blinkenlights (Re: New Book: Programming Embedded Systems in C and C++)
Brian Converse
> #Old-time hackers sometimes get nostalgic for blinkenlights because
> #they were so much more fun to look at than a blank panel....
Right as Sperry bought Varian's mini operation, they introduced a new line
of minicomputers (V77-600) that were microprogrammed. The front-panel
lights were driven by the microcode, and were originally used only in
diagnostic mode. After worried questions by potential customers, marketing
insisted that the machines be changed to put something, anything, in the
lights so that they would "blink". Accordingly, the machines were changed
to put the ID of the currently-executing task in the lights (and this
scheme might even work with some use today for the "150 Megacycle"
machines). Coming to later minis from the LINC and PDP-12, it annoyed me
to encounter "el cheapo" minis like the PDP-11 with only one switch
register bank and one set of lights. But it was definitely the end of an
era. It was also the beginning of an era when computers were more exposed
to ordinary citizens, who love to press switches, so economy and safety
combined to eliminate the console. Anyone who's had to toggle in a boot
loader into the default brain-dead PDP-8 and lower lifeforms of the PDP-11
family rejoiced at built in autoloaders: the lack of switches also meant
that one NEVER had to do that any more.
Maybe the radiation was a problem for military or TEMPEST situations with
sensitive receivers (or evil ones) nearby. But in commercial settings, the
only consideration was that any EMI not couple into adjacent circuits. At
the impedance levels used and with the thresholds involved, that wasn't
likely. With logic swings often in excess of 10 volts, an AM radio might
be useful for playing tunes via clever program loops, but the computer
wielded the inverse square law quite effectively against radio stations
that weren't literally next door. So much for recreational listening while
coding, and the whole idea of FCC emissions regulation was laughable.
While it's past the time that elaborate register displays outside of ICE
equipment are much use, in an embedded situation, flipping an LED can still
provide some help. First, we are very good at recognizing patterns, so if
there are LEDs to indicate program state, even subtle modulation, if
stable, provides a sensitive "health" indicator. An experienced user can
often spot problems: "it doesn't look like it usually does". One common
trick is to flip the LED when in the "idle" loop, which provides a telltale
of how many spare cycles are around. Although these days, everything wants
to be asleep and off when "idle", not sucking 10mA.<G>
--
Brian Converse
Exxon is about the 26th biggest institution on the planet, and that's
counting governments. . . . I think it galls them in their heart of
hearts that they're subject to the same justice system as everybody
else is"
- Brian O'Neill, lead attorney for Exxon Valdez claimants, quoted in LA Times.
Daniel E. Macks
: Schol-R-LEA wrote:
: > #
: > #Old-time hackers sometimes get nostalgic for blinkenlights because
: > #they were so much more fun to look at than a blank panel....
: While it's past the time that elaborate register displays outside of ICE
: equipment are much use, in an embedded situation, flipping an LED can still
: provide some help. First, we are very good at recognizing patterns, so if
: there are LEDs to indicate program state, even subtle modulation, if
: stable, provides a sensitive "health" indicator. An experienced user can
: often spot problems: "it doesn't look like it usually does".
*nods*. I'ven't a clue who cares any more what value is in the
accumulator of an Aspect processor, but the fact that the LED numbers
are a blur of changing values means it's working. And the fact that
various IRQs blink in a regularish pattern means there's more than
just an idle loop. If ADAKOS ever seems to hang, check the acc; if
there's actually a single number readable, time to reboot.
dan
--
Daniel Macks
dma...@a.chem.upenn.edu
dma...@netspace.org
http://www.netspace.org/~dmacks
Paul E. Bennett
bcon...@ids.net "Brian Converse" writes:
> Anyone who's had to toggle in a boot loader into the default
> brain-dead PDP-8 and lower lifeforms of the PDP-11 family rejoiced at
> built in autoloaders: the lack of switches also meant that one NEVER
> had to do that any more.
I've done that for Pr1me's as well. Not only boot-loaders but also short
test programmes to determine why the machine would not run the OS. Even
before doing this on Pr1me I had had practice on a 6800 system I built
and keyed the whole programme in this way (about 4k). It was quite a task
but satisfying once it was running.
If you want lights back, you'll have to build your own now.
--
Paul E. Bennett ................... <p...@tcontec.demon.co.uk>
Forth based HIDECS Consultancy
Tel: +44 (0)7971-620145
Going Forth Safely
Jens Kilian
> Maybe the radiation was a problem for military or TEMPEST situations with
> sensitive receivers (or evil ones) nearby. But in commercial settings, the
> only consideration was that any EMI not couple into adjacent circuits. At
> the impedance levels used and with the thresholds involved, that wasn't
> likely. With logic swings often in excess of 10 volts, an AM radio might
> be useful for playing tunes via clever program loops, but the computer
> wielded the inverse square law quite effectively against radio stations
> that weren't literally next door. So much for recreational listening while
> coding, and the whole idea of FCC emissions regulation was laughable.
> First, we are very good at recognizing patterns, so if
> there are LEDs to indicate program state, even subtle modulation, if
> stable, provides a sensitive "health" indicator. An experienced user can
> often spot problems: "it doesn't look like it usually does".
My first computer was a Sinclair ZX80, whose video output was driven by its
Z80 CPU. When a program was running, the display went dark.
The ONLY way to track what was going on was to put a radio beside the
computer and listen to the noise[1]. I still remember the lovely KA-CHUNK of
the floating-point trigonometry operations...
;-)
Jens.
[1] The ZX80 came in an unshielded plastic case.
--
mailto:j...@acm.org phone:+49-7031-14-7698 (HP TELNET 778-7698)
http://www.bawue.de/~jjk/ fax:+49-7031-14-7351
PGP: 06 04 1C 35 7B DC 1F 26 As the air to a bird, or the sea to a fish,
0x555DA8B5 BB A2 F0 66 77 75 E1 08 so is contempt to the contemptible. [Blake]
Stephen Maudsley
Jens Kilian <Jens_...@bbn.hp.com> wrote in message
news:FA2xt...@midway.uchicago.edu...
>My first computer was a Sinclair ZX80, whose video output was driven by its
>Z80 CPU. When a program was running, the display went dark.
>
>The ONLY way to track what was going on was to put a radio beside the
>computer and listen to the noise[1]. I still remember the lovely KA-CHUNK of
>the floating-point trigonometry operations...
When I started work there were remenants of a computer that used
acoustic mercury delay lines as the storage medium (a pipe containing
mercury with a loudspeaker at one end and a microphone at the other with
a read/modify/write circuit between) in one of the labs. When the
machine operated the piping hummed and you could hear different programs
operating by the different "song" they generated. It was said that the
programmers were hired on the basis of having absolute pitch as the only
way to debug programs.
It wasn't really an embedded computer but like most embedded computers
memory leaks were a health hazard.
--
Stephen Maudsley mailto:Stephen....@esgem.com
Esgem Limited: embedded system design http://www.esgem.com
Tel: +44-1453-521626 Mobile: +44-370-810991
Personal pages: http://www.esgem.com/people/Stephen.Maudsley
Peter Karlsson
Jens Kilian <Jens_...@bbn.hp.com> writes:
> My first computer was a Sinclair ZX80, whose video output was driven by its
> Z80 CPU. When a program was running, the display went dark.
> The ONLY way to track what was going on was to put a radio beside the
> computer and listen to the noise[1]. I still remember the lovely KA-CHUNK of
> the floating-point trigonometry operations...
My first computer was a Commodore 64, and I had the audio out connected
to my stereo. By turning up the volume on the stereo, I could trim the
azimuth angle on the C-64 cassette player by just listening to the
noise from it. I got quite good at it. :-)
--
Kind regards, Peter Karlsson.
Integration & Verification. Product development MDE.
Ericsson Microwave Systems AB. Base Station Division.
Office address: Fl=F6jelbergsgatan 2a, SE-431 84 M=F6lndal
Phone: +46 31 747 0406, Fax +46 31 706 7572
(drop your pants to reply by email)