On 1/20/04 10:58 AM, Claivaz Jean-Blaise <Jean-Blaise.Claivaz@...>
wrote:
> I'm looking for an open source portal that can federate multi-resources
> searches, e.g. retrieves datas from different databases through one request
> (OPAC, web-list, commercial DB...). It would be great if Eric has already
> planned the development of such a feature in his excellent MyLibrary software.
Claivaz, thank you for the endorsement. Your Euro's are in the mail. :-)
MyLibrary development
MyLibrary development is focusing on a number of things right now, listed in
no priority order:
1. We are in the process of re-writing the underlying Perl modules to be
more object oriented, and in the process we have adopted a development
process where we write the documentation first (the POD), write regression
tests second, write the module(s) third, and finally write the code.
"Thanks, Ed Summers!" The process is working quite well. Hats off to Rob Fox
who is doing most of the work. Instructions for downloading the software via
CVS can be found here:
http://dewey.library.nd.edu/mylibrary/cvs/
2. We are working on import routines. These routines will enable us to
read a valid RDF file, and import the data into MyLibrary. We expect data to
come in the form of MARC records from catalogs, translated into a flavor of
XML, and then inserted into MyLibrary. RDF will become the canonical file
format. Using XML provides more opportunities than MARC.
3. We have redesigned the database. The most significant change is the
implementation of resource->term->facet triples. MyLibrary will allow
libraries to define any number of facets denoting classifications. Examples
might include: subjects, audiences, types, or formats. Terms will be
individual headings within one or more facets. Examples might include:
Astronomy, Philosophy, Music, Freshman, Faculty, Senior Citizen,
Encyclopedias, Dictionaries, Maps, Books, Journals, or Data Sets. This
redesign was done because traditional library classification does not
provide enough granularity for individual recommendations of resources. BTW,
we have already written a script that should convert existing MyLibrary
implementations into this new database structure. A slightly dated and
difficult to read entity relationship diagram can be found here:
http://dewey.library.nd.edu/mylibrary/3.0/er-diagram.png
4. We are keen on syndicating MyLibrary content out of the library as XML
streams. This will enable MyLibrary content to be displayed in "portal"
applications beyond the library. Do date we can syndicate our content to SCT
Luminus and uPortal. The why's, wherefores, and links to demonstrations were
posted yesterday here:
http://dewey.library.nd.edu/mylibrary/mailing-list/2004/01/0001.html
Meta-search musings
Meta-search seems to be the Holy Grail we are all seeking these days. I
believe this is what Z30.50 was designed to do. At the same time I wonder
whether or not librarians have thought this one through. What content would
such a thing search? Where does this content come from? To what degree do
libraries design such tools for the lowest common denominator and at the
same time balance the needs of the expert searcher?
Traditional Z30.50 is too slow, too complicated, and too library-centric to
be a scalable solution. Maybe one of the Web services-based Z39.50
implementations will be of some use.
I think the real solution to the meta search problem is the acquisition of
the data itself, not just access to it. Just give me the data -- the
full-text content of the electronic journal or bibliographic index -- and I
will create one or more indexes of that content, and then I will combine
those indexes into a meta search tool. This way I will not have to create in
interface to other interfaces. Furthermore, such a solution is more in line
with one of the principles of librarianship, namely collection.
This being said, I too am seeking a meta-search solution, and I am hoping to
integrate it into MyLibrary.
--
Eric Lease Morgan
University Libraries of Notre Dame
(574) 631-8604
|