20 Sep
2013
20 Sep
'13
7:43 p.m.
On 9/20/2013 7:12 PM, Brent Hilpert wrote:
On 2013 Sep 20, at 5:57 PM, Josh Dersch wrote:
I recall it being a one-off, experimental sort of thing from the late
60's (where the thesis was something like "compilers and interpreters
are reallyinefficient and will never be as good as machine code, but
what if we did it in /hardware/, well, then...magic.").
I did find a reference to the Fairchild SYMBOL, which seems to be sort
of what I'm describing, but I still don't think it's what I was
originally remembering...
- Josh
So Iwas taking a closer look at the HP 1351A
"Graphics Generator" I
have. It's currently in operable condition, except that it won't
draw text (apparently there's something wrong with the character
generator, the :TX commands do nothing at all). I was surprised to
discover that the 1351 (c. 1981) doesn't containa microprocessor or
microcontroller of any kind -- all parsing of the command language it
supports is done in hardware (TTL). (The service manual contains this
gem: "The 1351A only accepts commands listed in this manual and in
the Operating and Programming manual. Any others, especially those
mentioned above will 'lock up' the 1351A such that it will have to be
re-initialized...")
At any rate, this reminded me of something that I'm prettysure I read
here on cctalk years back, but I can't seem to find any reference
toit anywhere; there was a computer designed at some university that
ran an interpreted language (I'm pretty sure it was BASIC), on the
metal-- that is, much like the 1351A and its vector description
language, this machine's hardware parsed BASIC program text and
executed it directly, rather than implementing some machine language.
Does this ring any bells, or am I insane? (It's quite possible that
both of these things are true...)
Perhaps you're thinking of the Wang 2200, which I'm led to understand
implemented BASIC in microcode; I think it may have been mentioned on
the list a few years ago.