- From: Leo Sauermann <leo@ist.org>
- Date: Thu, 22 May 2003 14:05:06 +0200
- To: "Rdf-Interest" <www-rdf-interest@w3.org>, <Patrick.Stickler@nokia.com>, "'Sherman Monroe'" <shermanmonroe@yahoo.com>, <zednenem@psualum.com>
I realised that we three are talking about connected issues so I dare to summarize the status. shunn me if i did wrong == Important == everything here was point to discussion. see older mails about: "Standard URI Set, and Resource Description Protocol (rdp://)" "URIQA" if you are interested: READ THE WHOLE MAIL ==The problem== There is a space of existing URIs that point to HTML pages that we access through HTTP. We call this "World Wide Web". A WWW URI points to a resource but when we resolve it through HTTP, we get a HTML page, thats ok for humans. To access RDF Triples, we want some "other system" that we can use to get the triples that describe a resource. Also, because in "Semantic Web" an URI can be any uri in any scheme. A simple problem is to get "rdf:type" of " A special problem was getting the "foaf:firstname" of "person@mail.com", where should we query. ==The Web approach== this is (what I understood as) the core of WWW * a URI should be a URL ! it points to "something". * a protocol is used to get the "something" * "something" is marked up Markup is solved: RDF/XML and other notations. fine. So we are searching for a good URI scheme and Protocol. == URIs AND URLS== -Patrick: use the existing urls URIQA- the header is intended for interaction with the web authority of the URI itself. -Sherman Semantic Web consists of Concepts- But in the semantic space, URI's can point to anything, not just documents. The machine can't use the URI to fetch a document, because there are no documents in the semantic space. So if there are no documents in the semantic space, then what does the semantic space consist of? It consist, on the most rudimentary level, only of a bunch of symbols that are used to dereference concepts. -Sherman: Ontologies- If someone/somegroup creates an ontology, then decides to discontinue maintining it, the ontology's URIs can still remain and flourish in the semantic space. Leo: actually i have to quite search for ontologies noted in RDFS because the namespaces normally point to nowhere, sometimes I just write the RDFS myself. -Sherman: RDP- It solves the issue of what should a URI return in a browser. This will once and for all place semantic resources in a space separate from the www. In this way, the semantic URI is viewed only as an atomic symbol that simply and unambigously "stands for" some concept or resource. -Leo use rdf to find documents- Humans use the rdf triples to FIND the documents readable representation and read them. -Leo: strange URLs and "RDF" as servername for resources- http://leo@rdf.leolize.it/outlook/appointment/12301928301823098123 this points to an outlook appointment of user Leo. The web server has an alias that catches the phrase "outlook" and hands the URL to an outlook adapter. the adapter parses and finds user leo. with this, you could even request HTML or RDF/XML content, with URIQA scheme. note also that the server name is "RDF" (could be virtual) -David: can use content negotiation- Second, I have to say that I'm not yet convinced by TBL's argument. It seems to me that I can easily assign a URI to my car (eg, <http://example.com/my/car>) and other URIs to documents describing my car (eg, <http://example.com/my/car.html>, <http://example.com/my/car.jpg>) and use content negotiation and the Content-Location header to return an appropriate representation if someone happened to dereference <http://example.com/my/car>. - XRI by OASIS - A new uri thing that conforms and can identify resources that are on more than one server. well, I didn't find anything about [XRI] on [OASIS] so I assume we have to wait a little till they reach their high goals. I think they are not really part of Semantic Web community, or ? == PROTOCOL == -David: use an encoding for complex queries- get information about <http://example.org/document#fragment> or <news:foo@bar>. I would rather use <http://sw.org?uri=(encoded_stuff)> than a straight URI with an extra header. -> -David: classes of web services- For a long term solution, what we would want is a way to describe a classes of web services. The class would define functions like concise_bounded_description(uri) and individual services would bind that to a particular set of URIs. -Leo: define a whole API- I will use the query functionality to connect all my desktop applications and will need a set of API functions, that can be accessed superfast with C or RMI. The api should be similiar to the web services. -Leo: build some modules- At the moment i am building modules for a server that handle the different urls. but this is based on ALIAS functionality, forwarding requests to a path to a module. -Patrick URIQA- Use the URIQA model on existing web servers. add "URI-Resolution-Mode: Description" to GET/PUT/DELETE requests answer with a "concise bounded description" Use URIQA as basis and extend it [URIQA] == CONTENT == -Patrick:Concise Bounded Descriptions- A concise bounded description of a resource is precisely that exhaustive body of knowledge known about a named resource which does not include any explicit knowledge about any other named resource. [URIQA] - EVERYBODY - Content should be some RDF Notation. Default RDF/XML == CONCLUSIONS == -UNIQUA- is based on existing structures. it uses a uri-resolution-mode:Description which is relatively easy to implement. It is a basic start that needs more complex ideas, too. - Seperate RDF space and WWW space - this was heavily discussed and has some drawbacks with protocol, ports, building new structures. == OPEN ISSUES == -TEST IT- We need some protocol & uri scheme that we can use to TEST our ideas on a WIDER scale, best would be to share our own information. alas, I don't own a web server so I will think about some trick. @Patrick: what useful Nokia information do you have for us that we can query? @Everybody: do you think about implementing UNIQA, what would you publish ? -URIQA- How would you decide that a server has URIQA and what will trigger you to make a GET/PUT/DELETE request with URI-Resolution-Mode: Description Do you have an idea how to manage the Website as a whole ? URLs on websites are in permanent change, what will be "best practice" to synchronize the Representation and Description ? -URIQA Implementation- how would an Apache2.0 module look like that does the URIQA stuff ? (patrick used TomCat & Jena) -EMAILs and STRANGE URIS- How to get the "Description/RDF" Triples about my email address "leo@ist.org" ? -COMPLEX QUERIES- Should we think about a protocol for "RDQL" or even "Daml Query DQL" ? yes, we should because the bigger index servers will do some inference work for us. If we can include the RDQL stuff, too, searches can be made distributed. -Concise Bounded Descriptions CBD- What about a less restricting definition of it ? Musicbrainz has already a running service that solved the problem with "depth", we could introduce a "level" of CBD: "CBD:plain" just the uri and connected unnamed uris "CBD:withLabels" get me at least the human readable labes of connected URIs "CBD:withSemantic" get a wider view of the resource that was defined by a human. (f.e. a CBD of a Music Album CD also retrieves the Artist and the Tracks) see [MUSIC] == References == [URIQA] http://sw.nokia.com/URIQA.html [MUSIC] http://www.musicbrainz.org/MM/index.html [Jena] http://www.hpl.hp.com/semweb/jena.htm [XRI] http://xml.coverpages.org/ni2003-01-08-a.html [OASIS] http://www.oasis-open.org/glossary/index.php
Received on Thursday, 22 May 2003 08:05:03 UTC