I think that if someday we have DSpace instances in the LOD cloud, we
are talking about multiple sites publishing a service and not just
one, likewise, we are using RDFS like DCMI elements, terms and types
and ORE resource maps. So, if someday in the future, one were to
update the cloud to include this, what would be its representation?
DSpace? DCMI? With usage of DCMI RDFS already in the cloud, should
that be a bubble in its own right?
-Mark
On Apr 3, 2008, at 9:18 AM, Hugh Glaser wrote:
(Thanks for adding the RKBExplorer stuff, Richard.)
With reference to size, which of course matters:
One of the nice things of Richard's cloud is that he does not get
pedantic
about exactly what a bubble means. So some of them are
straightforward LOD
sites; others are multiple sites, and still others are almost just
ontologies against which people are publishing linked data. This is
good,
because otherwise we would have long discussions about the
semantics of
bubbles and more worringly arcs!
But perhaps a little more meaning could be introduced to give a
sense to
casual observers (and others) that this is no just a collection of
27 (or
whatever) sites.
Would it be hard to make some of the bubbles (such as FOAF and
RKBExplorer)
clouds themselves, to indicate this?
I rather like the idea that the LOD cloud has become a cloud of
clouds.
Best
Hugh
On 01/04/2008 23:15, "Uldis Bojars" <[EMAIL PROTECTED]> wrote:
On Tue, Apr 1, 2008 at 12:32 PM, Richard Cyganiak
<[EMAIL PROTECTED]> wrote:
Specify the amount of data ( resources or triples ).
Individual and aggregates ( per type? )
Strength is in the numbers!
I agree that a vocabulary for describing datasets would be a good
thing. And
keeping track of and publishing numbers about the amount of data
would also
be good. I'm afraid I don't have the bandwidth to do any of those
things at
the moment, but if anyone has some spare cycles and wants to
chronicle the
project's growth in a more quantitative way, that would be great.
The chart would look more scary if it had some indicator of the
amount
of knowledge it conveys!
Scarier than a bunch of circles with funny acronyms that don't mean
anything to most people.
That's a very good point.
The beauty of the current picture (thanks, Richard!) is in its
simplicity. Anyone can look at it and say: "I understand this. Linked
data is a great idea.". Cluttering figure with numbers may look scary
but will this "scary-ness" help or defeat the purpose of the
figure? I
am afraid it will be the later for many. Think iPhone versus more
complex-looking (but less successful) devices.
Having said that, if someone collected together and kept track of
numbers, that would be a great resource. Our colleague Sheila [1] has
done some work on mapping ontologies / namespaces on the Semantic
Web.
While her work does not map 1:1 and is at a finer-grained level,
perhaps it can feed into work of analyzing linked data usage on the
web if someone is doing that. (Which might not be that trivial of a
task, unless someone already have the numbers at hand)
[1] http://www.deri.ie/about/team/member/sheila_kinsella/
P.S. Just to reiterate: not against quantitative indication of the
amount of linked data, but would keep things simple and put them in a
separate table / figure.
Uldis
[ http://captsolo.net/info/ ]
