[SCC_Active_Members] Software Archive Problem Statements?

[Tim Shoppa]

Eric Petrich <epetrich_chm at make-do.org> wrote:
> Do we need both short-term and long-term solutions? Can we (or the
> museum staff) come up with a "problem statement" regarding how to get
> software archived with enough metadata and incorruptibility to allow
> formal accessions later on?
>
> My wife suggested burning stuff to CDs or DVDs as a way to ensure
> (albeit temporarily) that we have copies that can't be altered. That's
> an implementation without a complete understanding of the problem, but
> I'm trying to illustrate how we can have a temporary solution that
> wouldn't necessarily be part of the long-term plan.

If you are worried about deterioration of the media causing
corruption undetected by the CD-reader, or something like that, then
"md5sum" or some other high-reliability checksumming system
is perfect. You store the checksum outside the copy itself (maybe
on a stone tablet if you believe those are more reliable than
paper or CD's!)

This technique has other advantages too, especially
valuable for those who want to compare/sort/uniquely identify
large sets of items with small unique tokens. In fact, software
versioning systems internally use such identifiers as part
of their internal sorting.

As to some legal proof that the image you burned to CD is
actually a full and complete image of the original media for
all possible legal/judicial purposes, that's
harder, but if it's part of the requirement, you're gonna have
to cough up the resources to make the solution address it. This
problem is a million times harder than, say, electronic voting,
so budget time and effort accordingly :-).

Tim.