10 Feb
2023
10 Feb
'23
5:58 a.m.
That's all the more reason to be precise, with all the uncertainty there is at least
the fact that I can say this is what the documentarian in charge of the manual had
approved to live in /usr/man/u_man on the particular box they had hooked up to the
typesetter when they cut the plates that eventually printed the book I hold. No amount of
printed material will ever capture reality, but it does give a darn good window into it in
a more precise way.
On the note of precision, to add fuel to the USG fire, there is also the USG Program
Generic line, which has numbers distinct from this stuff too. So many branches...
- Matt G.
------- Original Message -------
On Thursday, February 9th, 2023 at 8:48 PM, George Michaelson <ggm(a)algebras.org>
wrote:
I hesitate to throw this into the mix, but I would
observe in those
days, people were a bit less formal about release versions and I have
a suspicion (but only a suspicion) that from time to time what shipped
on a 1200bpi tape was not some canonical 'this is the release as of 3
months ago' but more 'this is a sh scripted product of the checked out
state as I understood it, on the box I had available to me, to cut a
tape.
The tape had to have boot blocks up front. You got told to do mt fsf
stuff. So.. I accept some of the tape structure was a bit more
formally policed: giving people the wrong architecture and bootblocks
would be bad.
But when it comes to what was unpacked for runtime? I think it wasn't
quite as 'reproducible build' formal as it is now, for some people.
The version numbers were not mutable. What was stamped with them?
Perhaps it was.
G