[Dspace-general] Questions, Relevance

[David Provost]

For those on this list familiar with both DSpace and Simile
(http://simile.mit.edu/):

It's beginning to seem as though the Simile suite will either transcend the
capabilities of DSpace or simply render its functionality as limited and
unnecessary. I'm arriving at this conclusion after reading Chris Bizer's
meeting notes sent Thu 3/24/05 (pasted below), reading messages from the
Simile list, listening to others, and just trying to think this through in
general. 

Am I perceiving the evolution of these technologies correctly?

Here's how I've been looking at DSpace and Simile:

DSpace is an abstraction layer above a data store(s). Tools from Simile can
then be applied to this abstraction layer to annotate/tag, create views, and
ultimately weave together electronic artifacts no matter their type or
location. I also include programs in this notion of weaving (eventually, if
not now) that will allow individuals (in a corporate setting, authorized
users) to create and/or tailor the functionality they want or need.

Thanks,

David

____________________
David Provost
978-549-5356
www.davidprovost.com

================================

Chris Bizer's meeting notes, sent Thu Mar 24 '05, to the
general at simile.mit.edu mail list:

Hi Stefano and Emmanuel,

we had an interesting meeting with David Karger today, focusing mostly on
the intermediate format a Fresnel engine could produce.

Because Haystack supports editing data, they need the URIs of instances to
be passed to their widgets so that they can update the underlying RDF
repository if necessary.

This wasn't included in the intermediate format I proposed 2 days ago, but
is no problem to add.

This thought lead us into the direction, that it might be a good idea to
implement one Fresnel rendering engine that would provide all Fresnel
functionality and output an intermediate tree.

This tree could then be used by Longwell, Haystack, IsaViz and Cocoon to
render whatever they like. Meaning that the tools would share as much code
as possible. I'm not implementing, so it is not my business, but looks like
an interesting idea.

Seen on an abstract level (not XML), the tree would be a tree of RDF nodes,
where each node has a set of CSS styling instructions attached to it. Pretty
general and the kind of input all four tool need.

I will update the documentation and the RDFS vocabularies now and freeze the
stuff afterwards. I will be in New Jersey over Easter and will be back on
Wednesday afternoon. I guess that I'm online only occasionally till then.

David Karger will read the manual in detail until next Thursday and we will
meet with him then again to see if he found stuff that would make it
impossible for Haystack to use Fresnel. He and I don't think so, but we will
see.

I will fly back to Berlin next Saturday, so I will basically be at MIT only
two more days (next Thursday and Friday). Maybe I will spend some time on
defining such an intermediate language then, so just see how the Fresnel
issue list looks next week ;-)

Cheers

Chris