Institutional Cost of Creating/Maintaining an OA Repository

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Mon, 6 Mar 2006 22:56:00 +0000

Yet another reason -- apart from focus and urgency and immediate benefits
-- for modularizing institutional archiving needs and efforts, and
focusing on an immediate OA-specific solution for OA content (journal
articles) is that it is cheap, fast and simple (unlike a Digital Omnibus
Archive). Here are some recent estimates in pounds (for dollars, multiply
by 2) from the new Eprints Wiki:

    http://wiki.eprints.org/w/How_much_will_it_cost%3F
    
    How much will it cost?
    
    Estimating how much an eprints archive costs...
    
    Basically the problem is that (so far) many archives are set up as
    a personal mission of a single person, who devotes spare time to it
    a research project side-effect. That is the funding is to set up an
    eprints archive and then do research based on it.
    
    The main cost of setting up is staff time which is very hard to
    calculate but I would suggest as a minimum:
    
    1000 - Good PC
    
    0 - Software - it's all free, although getting Redhat Enterprise is
    probably worth the small cost.
    
    1000 - per/year backup provision and network connectivity (wild
    guess!)
    
    Staff time: I am assuming:
    
    Little or no quality checking performed by a human being on data being
    submitted
    
    The UNIX admin already knows how to run apache, perl, redhat linux,
    XML etc.
    
    The archive will run an (almost) default configuration of EPrints. The
    default is suitable for a university archive.
    
    Setup: One week of skilled admin time to get a machine installed with
    redhat linux, the required libraries installed, eprints installed and
    an archive set up with minimal customisations (maybe a new webpage
    1000
    
    Per Record: Once someone learns the basics, adding an eprint
    is about 15 minutes work [1], so cost depends on "value" of
    staff member doing the deposit.
    http://eprints.ecs.soton.ac.uk/10688/
    
    UK Estimate: ~ 8
    
    Most people ignore the cost-per-record as it's absorbed by many
    peoples time.
    
    Also usually the staff costs will be increased by many many many
    meetings. The above is a theoretical minimum, assuming they just
    mostly take the default options.
    
    CalTech Review and Cost Estimate (2001) [2]
    http://www.arl.org/sparc/pubs/enews/aug01.html#6
    
Received on Sun Mar 12 2006 - 02:25:41 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:48:15 GMT