Re: Manipulating C code at the AST level, in C

On 2026-08-17, c186282 <[email protected]> wrote:

>    The old boxes had to be kept within a narrow
>    temperature range. Mostly that's be 'cooler'
>    but in winter might be 'hotter'.

Not as narrow as many people thought.  I used to get
into "thermostat wars" with people who believed that
the purpose of air conditioning was not to make things
comfortable - for people or machines - but to make
things COLD.  As with anything else a moderate
approach worked quite well - but most people just
couldn't stand the thought of a machine room being
comfortable, so after I left they'd crank the
thermostat right back down again.

Most installations I worked in only ran the machines
during office hours, shutting them down at the end of
the day.  One Friday evening I shut down the machine
but forgot to shut off the air conditioning before
leaving.  Monday morning our breath was condensing
in the machine room.  The oil in the disk drives'
hydraulic actuators congealed, and the heads wouldn't
load until we left the disks spinning for the couple
of hours it took for the oil to warm up.

At another such shop, the machine was notoriously flaky.
When the CE pulled one of the main boards out of the CPU,
a VLSI chip fell out.  The constant thermal cycling had
caused the chips to literally walk out of their sockets.
Once we left the machine on 24/7, the problems went away.

-- 
/~\  Charlie Gibbs                  |  In this world there are
\ /  <[email protected]>      |  two kinds of people:
 X   I'm really at ac.dekanfrus     |  1. Those who can extrapolate
/ \  if you read it the right way.  |  from incomplete data.