Re: [comments] SPARQL Protocol against QA SpecGL ICS [OK?]

Le 05-10-13 à 17:39, Dan Connolly a écrit :
> The SPARQL protocol doesn't introduce any extension mechanisms,
> but it is layered on various extensible technologies.
> To say simply "It is not possible create an extension
> of the SPARQL protocol" would introduce more ambiguities than
> it would remove, I think. Some might read that as saying that
> a SPARQL service cannot export other WSDL interfaces, or that the
> HTTP protocol cannot be extended with new methods, or that no new
> RDF class/property URIs may be introduced into the web, or
> that no new URI schemes may be introduced into the web.

That's the usual confusion that people have with regards to  
extension. :) What you said is perfectly fine and as nothing to do  
with extensions but with the provision of the language. To choose an  
analogy which might be clearer:

in HTML 4.01, there's a feature defined by the "class" attribute. The  
class attribute helps create semantics when an HTML 4.01 element  
doesn't exist for it. For example, we could have

<div class="poem">
…
</div>

It's not an extension of HTML 4.01 but one's might want to create to  
mark up the semantics of a poem.

<poem>
…
</poem>

That could have been authorized or not in HTML 4.01, with a precise  
and defined extension mechanism. HTML 4.01 has chosen to not be  
extensible at all, then you can't do it. The specification even  
recommends to ignore any elements which are not part of the DTD.

On the other side, CSS 2.1 has a well defined mechanism to extend the  
list of properties of CSS without interfering with the provisions of  
the language.
     http://www.w3.org/TR/CSS21/syndata.html#q4

> If you have specific suggested text, we'll certainly consider it.

That is difficult because it's very dependent on the architecture of  
your technology. For example, we addressed it in QA Specification  
Guidelines with this text.

[[[
4.3 Specification Guidelines Extensibility

This specification is extensible.  That is, adding conformance- 
related information and structure to the document in ways beyond what  
is presented as Requirements in this specification, is allowed and  
encouraged.  Extensions to this specification must not contradict or  
negate the requirements in this specification.
]]] - http://www.w3.org/TR/qaframe-spec/#specgl-extensibility

> Or if you have a test case (sketch) that shows some possible
> extension that you have in mind that isn't already addressed
> by the specification (directly or indirectly via cited specs)
> we'll certainly look into that.
> Kendall, does any text regarding extensibility come to mind?


I would have been happy to define a text if I was a specialist of  
your technology which I'm obviously not.

If you and Kendall needs to more about Extensibility and Extensions,  
please see this short section
     http://www.w3.org/TR/qaframe-spec/#extensions


Best
-- 
Karl Dubost - http://www.w3.org/People/karl/
W3C Conformance Manager
*** Be Strict To Be Cool ***

Received on Tuesday, 18 October 2005 18:34:00 UTC