RE: DSpace container architectuer discussion

[Mick Bass]
how relationships among the information could/should be expressed
....
And here is the required context information, including links to the
diagrams that were being discussed

[wmj]

What kind of relationships is represented by black and red arcs in the
diagrams?
I am converting/rewriting such diagrams into another notations which
"demands" such spec.

TIA

WMJ

-----Original Message-----
From: www-rdf-dspace-request@w3.org
[mailto:www-rdf-dspace-request@w3.org]On Behalf Of Mick Bass
Sent: Wednesday, May 16, 2001 11:55 AM
To: www-rdf-dspace@w3.org
Subject: re: DSpace container architectuer discussion



>The attached text file includes a 16-May-2001 discussion regarding the
>DSpace container architecture, how information could/should be modelled in
>DSpace, and how relationships among the information could/should be
expressed.
>
>
>- Mick

And here is the required context information, including links to the
diagrams that were being discussed.

- Mick

>From: "Tansley, Robert" <Robert_Tansley@hplb.hpl.hp.com>
>To: "Mick Bass (E-mail)" <bass@MIT.EDU>,
>         "Peter Breton (E-mail)"
>         <pbreton@MIT.EDU>,
>         "David H Stuve (E-mail)" <dstuve@MIT.EDU>
>Subject: More diagrams!
>Date: Wed, 16 May 2001 14:55:12 +0100
>X-Mailer: Internet Mail Service (5.5.2650.21)
>
>Another couple of diagrams, with a new idea on how to handle the HTML (and
>other type of file) bundle:
>
><http://www.mit.edu/~rtansley/scrapbook/object-models/objmodel-2001-05-16-0
1.gif>
>
>A slightly altered object model for the same example record as before. The
>individual bit-streams constituting the HTML are stored as separate objects
>(but "belonging" or "contained within" a bundle object.) Another bit-stream
>within the bundle holds a map of bit-stream IDs and corresponding filenames
>in the context of the HTML bundle. This is system-generated information -
>perhaps the system should have an explicit distinction between this and
>submitted data; that could be as simple as a flag somewhere.
>
>It should be noted that those types on the right-hand side are first-class
>objects - they'd have a reference to representation information, as well as
>being used to determine which storage mechanism to use (etc.)
>
>That "HTML bundle metadata" could in fact be a more generalised "bundle
>metadata" object, with the information about the bundle actually being HTML
>indicated by the type of the bundle object.
>
>
><http://www.mit.edu/~rtansley/scrapbook/object-models/objmodel-2001-05-16-0
2.gif>
>
>As the above example, except there is no intermediate "bundle" object. In
>this case, the extraction of the HTML requires more effort on the part of
>the disseminating process/service.
>
>Rob.
>
>--
>  Robert Tansley
>  Hewlett Packard Laboratories Bristol
>  External: +44 (0)117 3128211
>  Telnet: 312-8211

=============================================
Mick Bass, Sloan MOT 2000

R&D Project Manager, Hewlett-Packard Company
Building 10-500 MIT, 77 Massachusetts Avenue
Cambridge, MA 02139-4307

617.253.6617 office    617.452.3000 fax
617.899.3938 mobile    617.627.9694 residence
bass@alum.mit.edu      mick_bass@hp.com
=============================================

Received on Wednesday, 16 May 2001 22:53:44 UTC