Re: PSA: Clock drift and pkgin

>> But a drift of a couple of % would not be unexpected.
> Um, yes, it would be unexpected.
> The crystal spec in that era was 0.01%, [...]

> If you see a couple % drift, that's definitely not the crystal.

I'm not so sure.

My experience is that older machines keep worse time than newer ones,
even high-quality older machines and cheapo newer ones.  I suspect that
crystal frequencies drift with time (or perhaps with something that
tends to correlate positively with time, such as wide temperature
ranges in storage).

And, whatever the cause, I don't find a couple of percent drift
surprising in practice.  I once had a machine (I forget what kind) that
had really good timekeeping, just not at the frequency it was supposed
to - exactly what I would expect if the crystal were something like
5-10 percent out of spec.  I had to manually fudge NTP's drift value to
something _way_ out of the normal drift range, but, once I found a
close-enough starting value that NTP settled down and refined the drift
setting, it kept good solid time.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B