- From: Norman Gray <norman@astro.gla.ac.uk>
- Date: Mon, 8 Mar 2010 16:20:10 +0000
- To: Andy Powell <andy.powell@eduserv.org.uk>
- Cc: 'Kingsley Idehen' <kidehen@openlinksw.com>, 'Linked Data community' <public-lod@w3.org>
Andy, hello. On 2010 Mar 8, at 15:45, Andy Powell wrote: >> I summarise it as "it's OK >> to have both" > > Hmmm... the paper ends with: > > "HTTP URIs with specialized prefixes provide greater capability than URIs based on new URI schemes or URN sub-schemes in virtually all cases. Furthermore, such HTTP URIs seem better equipped to survive the test of time than URIs based on new URI schemes or URN sub-schemes." > > which hardly seems like a ringing endorsement of "it's OK to have both"? I summarise it like that -- David Booth may or may not endorse that! In fact, the paragraph just before the one you quoted says "Are these differences important enough in practice to warrant creating a new URI scheme or URN sub-scheme? In my opinion, no", so I'm rather reading against the text, here. There are advantages to new URI schemes. This community doesn't need persuading that those advantages are outweighed by HTTP-based names, on the come-all-ye World Wide web. That doesn't annul all the new-scheme advantages, however, and what Booth's paper does (for me) is to indicate that here is a mechanism which allows a subcommunity to potentially get the benefits of new-scheme URIs without having to throw away the much greater benefits of HTTP names (this 'subcommunity' remark is what I meant, in the message to Kingsley, about there being multiple boundaries, each with an inside and outside). People seem to get terribly upset about all-HTTP or new-scheme URIs, in the same way people used to get terribly upset about XML syntax. But just as XML isn't really about pointy brackets, identifiers are not really about punctuation. People needn't obsess. All the best, Norman -- Norman Gray : http://nxg.me.uk
Received on Monday, 8 March 2010 16:20:41 UTC