DefiningImageAccess/Resource/LinkedData
From ImageWeb
| Defining Image Access | DefiningImageAccess/RelatedWork |
Linked Data
One of Tim Berners-Lee's excellent series of notes about the rationale and design of the Web.
This piece explores linking of data on the web, and proposes four simple rules to make it work (better):
Like the web of hypertext, the web of data is constructed with documents on the web. However, unlike the web of hypertext, where links are relationships anchors in hypertext documents written in HTML, for data they links between arbitrary things described by RDF,. The URIs identify any kind of object or concept. But for HTML or RDF, the same expectations apply to make the web grow:
Simple. In fact, though, a surprising amount of data isn't linked in 2006, because of problems with one or more of the steps. This article discusses solutions to these problems, details of implementation, and factors affecting choices about how you publish your data. |
Go to the page for more!
I think the second rule deserves a little more evaluation. I do not, overall, dissent from the principle as stated, but the analysis (as I have read it) does not fully take into account aspects of social expectation. TBL himself has made efforts to point out that URI stability is a social, not technical, function (http://www.w3.org/Provider/Style/URI).
For many people, especially technical people whose influence within an organization isn't always noticeable, http: URIs are subject to whimsical change. The set of URIs based on a company's domain name tends to be property of the marketing department, and liable to reorganization with every rebranding exercise. Even though there are alternatives (purl.org, w3.org, etc.), the expectation that people have of HTTP URI stability is low. This may change given enough years of education and good practice, but I'm not holding my breath. I think the great advantage of URNs, DOIs, LSIDs and similar schemes is that, especially within certain communities, they convey to the people involved certain expectations of stability and authority that are highly valued.
So how to bridge the gap to achieve Tim Berners-Lee's vision of linked and accessible data on the web? I don't have any complete answers, but some partial ideas, all flawed:
- IETF's work on DDDS (RFC 3401, RFC 3402, RFC 3403, RFC 3404) is one approach using DNS, but not widely used as far as I know.
- Google the URN - may work for people, but isn't really a machine usable solution for the Semantic Web.
- Metadata schema registry lookup: there's no clear path of authority.
- Use a standard mapping from (say) URN to HTTP URIs within some stable and neutral domain: again, no clear authority, yet it starts to take us towards needing a centrally administered facility to support URNs. (I suppose a regexp-based mapping served as text from the IETF URN web pages wouldn't add too much burden.)
- Add support for known schemes into browsers, web libraries, etc. Even if it might conceivably happen for very popular schemes, but this isn't a general solution.
How are we to address this in our data webs work? My argument above suggests that non-HTTP names are particularly popular within particular communities of interest who happen to use the web for their work. Communities that span all kinds of web user tend to use HTTP: anyway (cf. FOAF, Dublin Core, etc.), often using services like purl.org. As it happens, a data web is intended to target a particular community, so one reasonable function of a data web might be to provide some kind of resolution for non-HTTP URIs used by that community so that useful information can be retrieved on the web. It's not a perfect solution, but I think it's in keeping with our philosophy of data webs being a small step toward Semantic Web Nirvana, and at least suggests a migration path for our interested communities.
