Print

Print


Eric Lease Morgan wrote:

> The reason I think this is interesting is in the light of library
> preservation. Backing things up to tape is not archiving -- it's
> backing up. Putting things on CD is close to archiving, but the
> archivist needs to think about moving the data forward to newer mediums
> as older mediums become obsolete.
>
> The idea of using version control software to duplicate a person's data
> over many computers does really not do the archival thing, but it does
> to the preservation thing. One of the best ways to preserve a document
> is to duplicate it many times and put those duplicates in many places.
>
> Maybe us here in Library Land can figure out a way to use SVN to mirror
> copies of our websites or other digital information. Hmmm...

As an avid CVS user, i keep items like my resume as well as my code in
CVS ( a precursor to SVN ) and it is very usefull for managing my
ongoing and evolving work.  However I don't think that SVN or CVS is
usfull for purely backing up archival data, because the repository
itself, just like any other file, can be lost due to hard disk failures,
etc.  Most users of CVS or SVN backup their repository to multiple
locations (mirroring) or backing up to tape or other media.  The feature
that CVS/SVN offer is an easy way to access the files that are "tagged"
as a certain version at a specific time from remorte or local
locations.  Meaning, i can at some point in time, tag all of my files as
say version 1.5 and then at any point in the future I can pull all of
the files that are apart of my "project" at version 1.5.  This is
extremely useful for ongoing evolving work.  Since an archive doesn't
change, I would think that something like a mirrored file server or a
raid disk array may be the best practice in securing longevity of your
digital content.  In a mirror, if one server goes down, who cares, you
have others to take it's place.  Same thing for a mirrored disk array,
if one disk crashes, you have others to take its place.  Considering you
really wouldn't need the features that CVS/SVN offer, it would just be a
waist to use it, especially if CVS/SVN becomes obsolete in the future
when you go to pull some data out of the archive in 10 years.  I think
that relying on a software for your backups is more of a maintenance
nightmare rather than just keeping your files in their original form on
a server somewhere.

One thing for you to look into that might interest you is the pgfs
(PostgreSQL file system).  PGFS is a filesystem based on a sql database
that has built in versioning, etc.
http://www.linuxjournal.com/article/1383