What is MPRESS?
MPRESS is a concept/installation to provide quality indexing of mathematical preprints (servers). It is in itself operated in a distributed way.
MPRESS improves access to the full texts of preprints in mathematics by means of MetaData and provides comprehensive and easily searchable information on the preprints available.
MPRESS is an index of mathematical preprints and not a full text archive.
MPRESS - supported by the Math-Net Project (1997 - 1999) (M. Grötschel) - started under the auspices of the European Mathematical Society. The CEIC (Committee on Electronic Information and Communication), established by the IMU, recently has formed a subcommittee to support its further development.
What is MetaData?
MetaData is ``data about data''. A bibliographic record is an example of MetaData. MetaData are particularly useful when coded in a machine ``understandable'' form. MPRESS therefore supports the ideas of DublinCore.
Who can join:
MPRESS is open to mathematical research institutions/individuals which make(s) mathematical preprints freely available electronically on the Web and complie(s) with a modest standardization of MetaData.
MetaData should be produced by the author him/herself, because they should contain an abstract and MSC classification.
When using the Mathematics MetaData Markup http://www.mathematik.uni-osnabrueck.de/projects/META/ just use ``save as'' to download the MetaData file you created. It will contain the required coding.
We suggest to download the interface itself from ftp://ftp.mathematik.uni-osnabrueck.de/pub/MMM/ and to install it at your site.
The easiest way is to ask MPRESS to gather the MetaData files/preprints you created from your WWW Site(s). Therefore please send e-mail to MPRESS explaining where you store the MetaData files /preprints.
Departments/Individuals could for instance store their preprints in a
directory called /preprints/ .
The corresponding MetaData-files then should be found in a subdirectory
/preprints/shadows/ (as suggested in [HM, 1996], pg 16).
A simple naming
convention -faciliating maintenance- will call the DublinCore MetaData shadow of potatoe.ps by potatoe.DC.html.
You can - and depending on your local environment you perhaps have to -
procede differently.
The emphasis is on ``explain''.
Optional:
Math societies or major preprint
servers may want to install indexing software underlying MPRESS
by themselves and export the generated
index to MPRESS. In case you want to contribute in this way also,
we will supply further details.
Some information is contained in [HM 1996], pgs 3-14
(up-dated links at Harvest Sources).
The architecture of MPRESS
The most important feature of the MPRESS system is, that it is decentralized. That means the preprints themselves reside on the server their authors want them to. MPRESS (or a substructure) retrieves its content - summaries of MetaData/preprints - by gathering MetaData/preprints from (for instance) departmental servers or subject oriented preprints servers.
After collecting data MPRESS automatically creates an index. This index can be used to search for preprints and to retrieve them from the servers where they reside. The index is automatically kept up to date: Suppose Metadata/preprint become removed from or changed on a server: MPRESS will realize that and deletes/changes the respective entry (with a little delay). The software used is based on Harvest.