On Mar 28, 2006, at 5:33 PM, K.G. Schneider wrote:
> Wearing my LII hat (http://lii.org ) I have been approached by a
> library--and had suggestions on our user survey--for something I've
> wanted
> to offer from LII as an added value service (as in, we do it and
> you pay for
> it) but wasn't able to articulate very well either in execution or in
> technology. One survey response that I just read said:
>
> "Develop web services (accessible by subscription) to allow a
> developer to
> include some of the LII in an application."
Granted, I do not know all the details, and I'm sure I am missing
parts of the picture, but that being said I suggest you implement an
OAI interface to your content and restrict access to the interface
based on IP addresses.
Your content is saved in a MySQL database. Cool. You do I/O (input/
output) against the database using a content management system. Fine.
Install Perl on your machine and write a program that creates robust
OAI XML files using Dublin Core. Save these files to your file system
and put them in subdirectories denoting OAI sets. Download and
configure (there isn't much installation necessary) a suite of Perl
modules called XMLFile. [1] There is really only one file in the
suite to edit. Finally, restrict access to the set of OAI files by IP
address and let your subscribers harvest your metadata. Once
harvested the secondary websites will be able to incorporate your
content into their site as they see fit.
Alternative approaches/Web services to use might be SRU, RSS, or
Amazon's OpenSearch. I would stay away from searching technologies
though because they will be a constant drain on your resources. RSS
is not necessarily robust enough to adequately describe your content.
--
Eric Lease Morgan
University Libraries of Notre Dame
[1] http://www.dlib.vt.edu/projects/OAI/software/xmlfile/xmlfile.html
|