Subject: Re: clisp, clsql and foreign functions: problems in windows executables
From: (Rob Warnock)
Date: Sun, 11 Feb 2007 21:30:43 -0600
Newsgroups: comp.lang.lisp
Message-ID: <>
Tim Bradshaw  <> wrote:
| Michael <> said:
| > You know, some people actually do have lots of time but not
| > lots of dollars. I know that is hard for you to grasp, but
| > it is true more often than you apparently think.
| We call those people "unemployed" don't we?

Or "students", which certainly includes many/most "grad students".

Or even "staff" at some educational institutions. I know that when
I was a staff techie in an NMR research group at Emory U. in the 70's
that we always had plenty of time, and usually enough people, but
very little free cash. [The grant money mostly went for the people,
who were paid very little in any case.] So we employed a *lot* of
Wozniak-style bit-twiddley "cleverness" to build add-ons to the
group's DEC PDP-10 that we couldn't afford to buy:

- A home-brew "bus-converter" that let us use cheap, high-density
  (for the day) TTL gates to build interfaces instead of the low-
  desity, expensive {R,S,T,W}-series cards you were supposed to use
  to do I/O with the PDP-10. That, in turn, let us build...

- A 36-port serial interface implemented with only 36 input gates and
  36 output registers, the serialization/deserialization being done
  with "software bit-banging" using a kind of bit-parallel "SIMD"
  style which cost no more than doing a single port. [Only burned
  1.6% of the CPU for all 36 lines when idle.]

- A Calcomp plotter interface for under $20 instead of however many
  $1000's DEC wanted for one.

- An MDS line printer interface for under $20 instead of however many
  $1000's DEC wanted for one.

- Several other interfaces to various bits of laboratory equipment
  for $20-50 each, rather than the $500 to $10,000 prices of the
  commercial equivalents.

As The Woz himself has been saying during his recent book tour
[paraphrased], everybody should experience at least one time in
their lives when their technical work is *severely* constrained
by cost, or memory size, or a miniscule CPU, or some critical
resource *other* than human time. One learns a quite a lot by
having to make do with very little.


p.s. I learned to "think small" on an LGP-30 and a PDP-8 [and
carried those lessons over to the PDP-10 & PDP-11 & MC68000],
but these days I would suggest to new programmers that they try
their hands at programming some model of PIC (e.g., a 16F57 or
a 16F6xx), *in assembler*, to learn how to accomplish a lot with
little. Despite their limitations, they're plenty fast to run
all kinds of fun toys.

Plus, the PICs are *CHEAP*, only a buck or two each, and you can
build a serial programmer for the 16Fxxx parts that works off the
parallel port of your laptop for just few bucks in parts. Numerous
assemblers that run on Windows or Linux/Unix are freely available.

[Or you can buy the PICkit 2 Starter Kit from Microchip for ~$50,
which includes as USB-attached programmer, a demo board (with LEDs,
a pushbutton, and a pot), a PIC16F690 PDIP, and software. But where's
the fun in that?!?  ;-}  ]

Rob Warnock			<>
627 26th Avenue			<URL:>
San Mateo, CA 94403		(650)572-2607