Tim Shoppa wrote:
Somebody else mentioned databases... properly done
this could be quite
handy. But the existing directory hierarchy (multi-level) really is more
amenable to the actual needs than any database system (which
unfortunately tend to "flatten" the structure).
Oh, I'd only forseen the use of a database as a repository for metadata:
description, name of publisher, source of original document and other useful
things - files still exist, at least as far as the user's concerned[1], in the
hierarchy as they are now.
[1] technically this doesn't *have* to be the case of course providing there's
always some programatical intermediate layer between the user and the data on
disk.
Flattening the structure is useful for some things,
but only if you
then "blow it back up" into a hierarchy. For example I could imagine a
directory of 3rd-party-DEC-compatible stuff that softlinks back into the
manufacturer's directories. This is NOT a flat structure... "A maze of
twisty passages all alike" is a better analogy!
Urgh :-)
I suspect the main problem with any of this is that bitsavers need mirror
sites - and any introduction of a database on the primary site means that the
mirror sites need to use the same database and the db contents need to be kept
in sync :(
cheers
Jules