18 Jun
2001
18 Jun
'01
2:24 p.m.
On Mon, 18 Jun 2001, Anthony Clifton - Retrocomputing.com wrote:
I'm pleased to announce that, very soon, I'll
be taking over the
retroarchive.org web site from Gene Buckle. The site includes
around 600 megs of files, primarily for CP/M, but increasingly
for other OS's and platforms as well.
In addition, I'm going to place the files on the regional hub so
that they can be retrieved via UUCP...to make it easier to get
them directly onto vintage machines without having to use
conversion software.
To do this, however, I may need to use another common archive
format. My concern is that alot of vintage machines may not be
able to handle zip files. Suggestions for a format?
Well, ideally it'd probably be nice to have everything archived in some
format which is most appropriate for whatever system on which the files
are intended to be read. I don't know how practical that is.
One line of thinking would be that if you have lots of disk space in the
archive, then just keep multiple copies of everything in a handful of
the most popular file archive formats.
Another line of thinking would be that if you have lots of computer cycles
to spare, then keep everything in some standard format and convert it to
the desired format requested by the end user. I think this would be a
great way to handle it assuming the conversions can be run, but it results
in a lot of work for the archive syste.
My thoughts would be to maintain a central repository of compactly
compressed files, then leverage UUCP's distributed nature to dilute the
computation intensive conversion problem. So let's say the central
archive server has the name "archive". You'd then setup other translation
hosts around the archive server to get the file to you in the format you
wanted. For example, if you wanted the .zip version of the archive, you'd
access the archive through the "transzip" host in the path
archive!transzip!foo!bar!myhost
If you wanted the .lzh version of the file, you'd use the path
archive!translzh!foo!bar!myhost
etc...
I don't know how you'd actually go about implementing this so that the
transzip and translzh hosts would automatically take the central archive
copy and convert it to the appropriate translated copy. But I think it
would be neat if it's possible. And of course I'd imagine it such that
anyone who wanted to setup a translation host of one type or another would
be able to do so. It's not like there'd be only one host for each archive
type. Maybe I'd have my own host named "brizip".
archive!foo!bar!brizip!myhost
You'd just have to drag the file through some system which would know how
to translate archives. Again, I've no idea how this would actually work,
but I think it's got a sort of intuitive feel to it.
-brian.