The precarious state of classic software and hardware preservation

[Warner Losh]

On Sat, Nov 20, 2021, 11:46 AM Paul Koning via cctalk <cctalk at classiccmp.org>
wrote:

>
>
> > On Nov 20, 2021, at 7:56 AM, Bill Gunshannon via cctalk <
> cctalk at classiccmp.org> wrote:
> >
> > On 11/19/21 9:33 PM, Steve Malikoff via cctalk wrote:
> >> Michael asked
> >>> What are we, as a community, to do to fix this and make sure that our
> >>> history stays peserved and isn't one bad day away from vanishing.
> >> Whenever some new vintage computing page appears I go to archive.org
> and submit the
> >> URL to them for the wayback machine. Often they've crawled it already,
> but not always
> >> so I think it does help.
> >
> > And what happens when you wake  up one morning to find archive.org is
> > gone, too?
> >
> > I remember hearing how the web was going to make everything perpetual.
> > And yet the list of things that have disappeared just gets longer and
> > longer.
>
> The web can make things perpetual if they are stored redundantly and in a
> distributed fashion.  But anything centralized is just as vulnerable as any
> centralized copy ever was, whether from risk of fire or flood, or
> abandonment.  And in the case of digital data the added complication is the
> loss of the necessary technology.
>
> The Long Now Foundation has done some good thinking about this; some
> others have as well.  There's a long-lived software achival concept with
> "Rosetta" in its name that might be worth more attention.
>
> One of the best reasons why GIT is good is that everyone has a full
> record.  With Subversion of CVS, you just have the one revision you have
> currently checked out, but GIT gives you (as the command indicates) a full
> clone of the entire repository.  So suppose that, say, GitHub suddenly goes
> belly up -- for any item in their collection that anyone anywhere has
> cloned, nothing is lost.  The same goes for any other project that uses
> GIT, such as some of what the Free Software Foundation keeps.
>

You can mirror both svn and CVS. Lots of projects that used them did that
so there are thousands of copies of FreeBSD's svn and CVS trees about.
Further, git let's you do a shallow clone, which tosses history.

What git does is allow for local changes and the default is to mirror the
whole thing.

Having recovered from a disk failure that took out a svn repo in the time
it took to copy the files back is nice, but it did take more planning. Git
has a lot of cool features (it's why FreeBSD moved to it), but you are
misstating the defaults are always on and that others can't do this.

What has helped a lot since source history was lost in the 80s and 90s is
the adaptation of SCM, more disciplined use of it and the expansion of
disks to reduce pressure to purge old stuff and the expectations that SCM
is available to all... the early cowboy days of Linux, FSF and others has
been replaced by better engineering practices. Things still can be lost,
but it's much harder now... until things are really old...


I'd say more of us need to be more paranoid about mirroring stuff.  Is
> there an archive.org mirror?  I don't know.  There are bitsavers mirrors,
> which is a good thing to know, but not all classic computer stuff is in
> bitsavers.  What about Wikipedia?  There's Infogalactic, but that's a fork,
> not a mirror.
>

Yea, I struggle with what to do with my rainbow material for the long haul.

Warner

>