I am with St. Mary’s University School of Law in San Antonio. The law library has been using Omeka to build a digital collection. A critical missing piece of this as a repository platform is that Omeka does not have a preservation layer.
I have the idea that a universal preservation layer can be made to overlay any content management system, by making a program to run check sums on a static file structure.
For any CMS it is possible to find a plug in to modify file structure into a static meaningful file structure. Daniel Berthereau’s Archive Repertory plug-in modifies the file structure in Omeka, so that files can be stored in meaningful directories and assigned meaningful names (the default is to rename files with a random alpha numeric string and put all in the same directory). So, technological barriers are low for getting a good file structure, no matter what repository platform or CMS you are using, no matter what repository platform or CMS you are using.
Rather than make a plug in to do digital preservation, it would be possible to code something that lies on top of a static file structure. It looks at the file structure, makes a list of files, runs a check sum for each file and stores that check sum along with the file location, then at a later date reruns the check sums and gives a report of which files have a different check sum (ie. gives a report of which files have changed from bit rot). This would not be like a plug in, which installs on a specific CMS and only works with that CMS, but instead is something that overlays files on any web server.
By overlaying any set of files on a web server, such a program could do digital preservation independent of repository platform or CMS. This would greatly lower technological barriers to smaller institutions performing digital preservation.
I am interested in talking with anyone who has worked with digital preservation, and would like to meet anyone who knows about coding automated check sum reports. I understand conceptually how to do this, but want to connect with more experienced coders who are interested in similar software.
Ultimately, I would like to code this alone or, better, with others.