- From: Story Henry <henry.story@bblfish.net>
- Date: Tue, 28 Oct 2008 10:36:23 +0100
- To: "Simone Onofri" <simone.onofri@gmail.com>
- Cc: "Semantic Web" <semantic-web@w3.org>
There is a lot of good reading here:
http://www.w3.org/TR/webarch/
One important element is URI opacity.
Also see "Cool URIs don't change"
http://www.w3.org/Provider/Style/URI
So URIs are not type oriented. They are names for things that have
types perhaps, but URI opacity suggests that one should not build URIs
and expect software to get semantics by analyzing the URI.
The representation returned by the resource can on the other hand
describe it. So
/person/simeoneonofri
would return
<> a foaf:PersonalProfileDocument;
foaf:primaryTopic :simon .
:simon a foaf:Person .
If you wanted to describe the /person/ collection you could have it
return
<> a foaf:Group ;
foaf:member <simononofri#simon> .
Or something like that. (Not sure if a foaf:Group can be a
foaf:Document)
Henry
On 28 Oct 2008, at 09:33, Simone Onofri wrote:
>
> Dear Semantic Folks,.
>
> I'm building a demo of an RDF-based - file system storage on the first
> time - application based on Tim's SemWebAppMetro [1]. So I would like
> to propose an update of the map linking place data to on Photos.
>
> Secondly I'm thinking about the queston of URI. I know this can be
> influenced by some factors and application itself but what is the best
> practice - in general and in this case in particular? Or there are
> some pointers about this?
>
> By the way URI can be type oriented
>
> /person/simoneonofri
> /calendar/simoneonofri
>
> or user oriented
>
> /simoneonofri/person
> /simoneonofri/calendar
>
> Cheers,
> Simone
>
>
> [1] http://www.w3.org/2005/Talks/0623-sb-IEEEStorConf/SemWebAppMetro.png
>
> --
> Simone Onofri
> http://www.siatec.net/
>
Received on Tuesday, 28 October 2008 09:37:18 UTC