[Metadata] Distributed vs. centralized metadata

[Eric Hellman <eric@hellman.net>]

There are (at least) two basic architectures for DOI metadata.

1. The publisher is responsible for providing and maintaining metadata,
either within the digital object itself, or as a separate object. A central
resolver stores only pointers to the metadata. Decentralized services
collect and process the metadata into applications.

2. A centralized service collects and maintains metadata, then provides the
metadata to centralized or decentralized metadata applications.

The recent metadata paper seems to come down on the side of centralized
metadata, but at the same time wants it to be "minimal". The question of
how to do applications which require more complex metadata is deferred to
the "DOIGenre" concept. This really needs to be fleshed out, because a
centralized core needs to work well with decentralized DOIgenres. (You
can't prevent decentralized genres from happening. If DOI acheives its goal
of facilitating e-commerce of digital objects, then sooner or later someone
is going to make money with an XXX-DOIGenre!)

Eric
Eric Hellman
Openly Informatics, Inc.
http://www.openly.com/           Tools for 21st Century Scholarly Publishing

------------------------------------------------------
Metadata maillist  -  Metadata@doi.org
http://www.doi.org/mailman/listinfo/metadata