Comments on: Best Practice Proposal for a DESCRIPTION Datastream

By: Re: [Fedora-commons-users] Where to store things you need to search?

Re: [Fedora-commons-users] Where to store things you need to search? — Mon, 22 Feb 2010 10:46:35 +0000

[...] Steve. I also ran across this quotation at http://dltj.org/article/description-datastream/ "working with Fedora's compulsory Dublin Core (DC) datastream started one thinking about the [...]

By: Sergio Berna

Sergio Berna — Tue, 17 Oct 2006 14:24:21 +0000

[quote post="109"]“to establish a practice of creating an in-line XML2 datastream with the label ‘DESCRIPTION’ that contains the primary descriptive metadata for each object.”[/quote]

Interesting proposal. Do you know of any recopilation project of these denominations of fedora inline datastreams?

In my case I have need for several descriptive datastreams. That need a richer vocabulary.

The first datastream would be the descriptive datastream, using the tag you propose.

Then would follow preservation datastreams, Rights management datastream, format especification datastream, relations management datastream and many more that maybe thinking a little ahead could be equally named such as your proposal implies.

By: the jester

the jester — Tue, 12 Sep 2006 16:32:12 +0000

[quote comment="3837"]At Indiana, we’re keeping as little as possible in Fedora’s default DC datastream, and lumping all other metadata into a METS document in a METDADATA datastream (philosophy, sample object).[/quote]

I would note for those that haven’t followed the “philosophy” link, that “as little as possible” in Indiana’s case is:

It currently includes only these items:
Title
PURL for the object (if this is an item-level object) in an Identifier field
Fedora PID for the object in an Identifier field
The “real” DC record (if present) will be in the [descriptive metadata] of the METS document, alongside any other descriptive metadata.
http://wiki.dlib.indiana.edu/confluence/pages/viewpage.action?pageId=441#FedoraMetadataStoragePhilosophy-DublinCore

That certainly is bare-bones, but it makes a great deal of sense.

Thanks for the comment and the links to the examples, Ryan!

The text was modified to remove a link to http://fedora.dlib.indiana.edu:9090/fedora/get/iudl:4420 on January 19th, 2011.

By: Ryan

Ryan — Tue, 12 Sep 2006 15:58:13 +0000

This is definitely a reasonable approach. Indiana, Tufts, and Virginia are all taking similar approaches. The only difference is in the details. At Indiana, we're keeping as little as possible in Fedora's default DC datastream, and lumping all other metadata into a METS document in a METDADATA datastream (philosophy, sample object). Tufts keeps a fairly complete record in the default DC, but the fully-complete records are in DCA-ADMIN and DCA-META (sample object). Virginia has their own metadata format, which can be found on their metadata site.

The text was modified to remove a link to http://fedora.dlib.indiana.edu:9090/fedora/get/iudl:4420 on January 19th, 2011.