[SCC_Active_Members] Software Archive Problem Statements?

[Larry Masinter]

> This is a HUGE difference from the known-correct assumptions
> archivists can make about digital content going into something
> like a document archive.

I'm not sure there's a presumption about 'known-correct'
for document archives, either. Original documents get
modified, stained, coffee spills, bleed-through --
in fact, the use cases you gave about a piece of paper
stuck to the original document, or better scanning technology --
apply to books in English as well as code in FORTRAN.

However, even if the nature and frequency of errors
are the same for code and text, the *requirements*
for correctness are higher for code that you expect
to run again than they are for text. And software
is republished, with changes, much more frequently
than texts usually are (with some notable exceptions,
e.g., newswire stories). 

The problem of dealing with odd or unusual data structures,
and unreliable recovery sounds to me that it might have some
common elements with software forensics...
http://www.google.com/search?q=software+forensics
had some interesting links.

Larry