Re: "How to Publish Linked Data" vs "Dereferencing HTTP URIs"

On 10 Jul 2009, at 07:07, Christopher St John wrote:
> I'm putting together a quick presentation on 303s and
> Linked Data for the local Dallas Semantic Web Meetup
> (it's part of a series of 10-minute lightning presentations)

Don't forget to mention hash URIs, they are often easier to deploy  
than 303-emitting URIs in practice.

In addition to what Chris said:

> Q2) The "Note" vs "Recommendation" thing is formal spec
> speak and may not mean what it appears to mean.
> Can someone comment? The wording "This is a draft
> document and may be updated, replaced or obsoleted
> by other documents at any time" in a Linked Data foundation
> document is fine if you're just experimenting, but could
> be alarming if you're considering, say, writing a commercial
> tool...

There are two main reasons for why there is no Linked Data W3C  

a) Usually, only new specifications get W3C Recommendation status.  
Linked Data does not involve any new specifications, but it's a  
collection of good practices for the use of existing specifications.  
Thus, it has to be described in lowly W3C Notes.

b) Recommendations can only be published by Working Groups chartered  
to do so. It's a *much* more heavyweight process than publishing Notes.

I would argue that commercial tools SHOULD be written before the  
technology has solidified. That way, your experience while developing  
the tool, and your customers' requirements, can influence the  
development of the technology. All successful technologies are  
evolving; most are just in the helpful habit of slapping a new version  
number on the technology every few years. Linked Data hasn't been  
around long enough to do that yet.


> I did a (relatively quick) archive search but I could have
> easily missed a discussion somewhere, apologies if this
> has already been gone over. And thanks for your patience
> with the geeky spec details.
> -cks
> -- 
> Christopher St. John

Received on Saturday, 11 July 2009 10:26:21 UTC