Re: Numbered backups

On 2/13/26 5:24 AM, Tim Bradshaw (as tfb at cley dot com) wrote:
> What I meant is that people romanticise how great those OSs were.  They were better in many (not all) ways than PDP-11 era Unix,

Oh come on. Genera was an entire universe away from the other systems of 
the day. I am not sure what "Lost Cause" is (or what your personal 
problem here is) , but I am unaware of anyone working on "bringing it 
back". It has been 40 years since then, and the only reasonable thing to 
do in that area would be to start over, incorporating some of the lost 
ideas and features. And I don't know of anyone even thinking about that, 
because essentially nobody even knows what was there. From the way you 
describe it, I am wondering if you even know.

> but they also ran on machines which cost large multiples of what a PDP-11 did

I could cite many reasons why the technology died, talking about 
business decisions at the companies for example. But the fundamental 
problem was that the technology was ahead of its time. Meanwhile, 
/fundamentally/ crappy behind-the-times technology became cheap and 
easy, and took over the world. And here we are.

We would like to think the best stuff is what wins, but even adjusting 
for the financial dimension of "best", it turns out that's Just Not How 
The World Works. And everyone assumes that the newer thing is better, 
because, shouldn't it be? I am sure you can think of plenty of things 
that disprove that.

And things to get lost, especially if they were only known to a few. The 
forces that drive technology can take attention away, and the wheel may 
or may not get re-invented for a very long time. When what really needed 
to happen was for that wheel to have (a) made life better all along and 
(b) continued to be improved upon.

What we had back then was just the beginning of things -- things which 
have mostly never happened. Given another half-century of development, 
the word should be in a very different place right now. (Just for one 
thing, literally trillions of dollars and untold lives have been wasted 
chasing security bugs and their fallout. Bugs which we made impossible 
to occur back in 1977. And that still goes on today.) Also, I can't 
click on shit. :)

I started hacking in 1973 and have lived through (and worked with) just 
about every facet of the computer world from the 1960s technology up to 
today. I am not a romantic, and when asked questions in that 
("favorites" or "best" or techno-religious) direction, my response is 
the same as always. To wit, "I hate them all. I can tell you which ones 
I hate less, though". And the most malleable of those, and 
correspondingly the most promising (and most advanced at the time) was 
the Symbolics Lisp Machine. That's not romance, that's just having lived 
in a better world that unfortunately almost nobody got to see.

This is going to be my only rant about this, because when I brought it 
up, I was making a joke, as indicated by the absurd suggestion of using 
1960s-1970s technology that is no longer around, and the smiey emoji on 
it. (But since someone wanted to take it seriously, you get this rant.)

As for the complaint that Unix does support multiple file systems, and 
that some of them have a version control system in them, I will offer 
the following observation. Nobody is using those file systems, at least 
not anybody in this conversation. I would encourage people to check 
those out and see if they solve the desired workflow requirements in 
terms of functionality. I suspect that they do not meet the 
requirements. Because otherwise, everybody would be doing that. And 
clearly they are not. To the point of, for practical purposes, "NOBODY 
does that".

What would really be nice is a file system that actually had the desired 
features, and that worked. In which case it would be the standard file 
system. There are really several use cases being intermingled in this 
argument. There are at least two kinds of "version control" implied. And 
there are multiple ways to implement them and their user interfaces. 
There are file systems that do some of it, but I don't think the 
combination that we're looking for actually exists at present. Otherwise 
everyone would be using it and none of this Emacs version numbering 
crappy hack would have even ever have come up.

Not as elegant as having the operating system take care of it for you, 
but I do have to wonder if this could not be solved with git. Maybe all 
that's needed is a good user interface for this in Emacs. Too bad that 
GNU Emacs is an impovershed crappy hack of a 1970 version of Lisp and 
ignored 10-15 years of Lisp and Emacs development at the time. Even 
though the author was fully versed and contributed to that technology. 
But was forced to use 1960s (C) era technology to implement it, which 
made doing a good job too hard. Making it so hard to build sich an 
interface that it has never (and may never) be done. But fuck, at least 
it was CHEAP. So I guess that's better.

Fuck.





Today there is no issue whatsoever and a cheap Lisp Machine (which would 
be good for more than Lisp, of course) could be created. B