Re: section 2. Adding GRDDL to well-formed XML for review

On Tue, 17 Oct 2006, Murray Maloney wrote:
> As a matter of opinion, I think that it would be distracting to include a
> discussion of why we didn't follow an alternate design route. The spec
> should expose those design routes we have chosen. Otherwise the spec
> will be filled with a lot of verbiage that it doesn't need.

On the contrary, I don't think there is a more appropriate place to spell 
out this distinction than in the specification.  GRDDL suggests an 
alternative route that is more robust for sure, but the GRDDL 
mechanism only accounts for *one* particular goal in transforming content.

XSLT can be used (in very similar fashion) to generate other artifacts 
(from the same source document) for very different ends and I think you do 
more harm than good in not spelling out this distinction *especially* when 
you consider how ubiquitous support for xml processing PIs is within 
browsers.

What I'm suggesting isn't so much a justification of not 
using PIs but rather being clear about how the GRDDL mechanisms relates to 
similar transformation mechanisms to give an appropriate context for 
(web client) implementors who will undoubtably have to reconcile this 
distinction.

> If you suggesting that we add PIs as an additional mechanism for either
> identifying the GRDDL namespace or for identifying GRDDL links,
> then I think you face an uphill battle against the Web Architecture.

I'm not suggesting accomodating PIs within GRDDL, but giving the 
appropriate context for implementors who have will have to contend with 
the various, automated ways XML content will be processed within their 
architecture - PI's being the most prominent example.

Chimezie Ogbuji
Lead Systems Analyst
Thoracic and Cardiovascular Surgery
Cleveland Clinic Foundation
9500 Euclid Avenue/ W26
Cleveland, Ohio 44195
Office: (216)444-8593
ogbujic@ccf.org

Received on Tuesday, 17 October 2006 18:46:57 UTC