RE: [w3c sml] section "resolving an SML reference" section missing.

I viewed it as a kind-of duck/rabbit statement (I'm sure you know that
old visual switch trick).  The author might have intended the statement
to be read as a rabbit; I read it as a duck.

 

I would accept the "rabbit" version of the statement as you have stated
it below :-).

 

Kirk Wilson, Ph.D.
Research Staff Member

CA Labs

603 823-7146

 

________________________________

From: Smith, Virginia (HP Software) [mailto:virginia.smith@hp.com] 
Sent: Wednesday, December 05, 2007 1:49 PM
To: Wilson, Kirk D; Sandy Gao
Cc: public-sml@w3.org
Subject: RE: [w3c sml] section "resolving an SML reference" section
missing.

 

That is quite interesting reasoning, Kirk.  :-)

 

I think it means to say:

 

A conforming model validator MUST support SML references as defined in
this specification.

 

--

ginny

 

________________________________

From: Wilson, Kirk D [mailto:Kirk.Wilson@ca.com] 
Sent: Wednesday, December 05, 2007 10:20 AM
To: Sandy Gao; Smith, Virginia (HP Software)
Cc: public-sml@w3.org
Subject: RE: [w3c sml] section "resolving an SML reference" section
missing.

What exactly does the statement "References MUST be supported by model
validators that conform to this specification" (section 4) mean?

 

In the context, I would take this statement to mean that every reference
must be supported by at least two model validators.  In other words, to
use just an SML URI scheme in a reference, the URI scheme MUST
(obligatorily) be supported by at least validators "out there" in the
market place.  If there is only one URI scheme validator "out there",
then must I include another scheme for which a validator exists in order
to meet this requirement on the reference itself?  (Ok, I'm being a
little ridiculous on this last question, but it is to point out that the
statement could say something radically different than what was intended
if read literally.)

 

Is the intention to something more like:  Processing SML references
SHOULD employ a model validator that conforms to this specification - ??


 

Kirk Wilson, Ph.D.
Research Staff Member

CA Labs

603 823-7146

 

________________________________

From: public-sml-request@w3.org [mailto:public-sml-request@w3.org] On
Behalf Of Sandy Gao
Sent: Wednesday, December 05, 2007 11:44 AM
To: Smith, Virginia (HP Software)
Cc: public-sml@w3.org
Subject: Re: [w3c sml] section "resolving an SML reference" section
missing.

 


Thanks Ginny for performing this extensive analysis. I think Valentina
and I reached the same conclusion when we were discussing 5291. 

I'd also like to keep the current organization. "Rules" are generally
easier to understand than "steps". 

I can also see how the steps could help people, especially implementers.
Also the steps helped us (the WG) to ensure that we got what we wanted. 

Maybe a compromise could be to keep the current organization, and
include the steps as a non-normative section? Or the other way around
(rules non-normative; steps normative), but this could be much harder to
do. 

So I can see 3 options: 
1 Rules only 
2 Both rules and steps, with steps non-normative 
3 Both rules and steps, with rules non-normative 

No strong preference.

BTW, a purely editorial comment. It seems that the organization of 4.2.1
and 4.2.2 looks a bit awkward. The current 4.2.1 and 4.2.2 overlap a
little bit; and 4.2.1 also depends on the second half of 4.2.2. Would it
help at all to have: 

4.2.1 At most one target 
4.2.2 Consistent Reference Schemes 
4.2.3 Identical Targets 

where 4.2.1 and 4.2.2 talk about different aspects of the "one target"
rule; both of them need the "identity" definition for targets, which is
defined in 4.2.3? 

Thanks,
Sandy Gao
XML Technologies, IBM Canada
Editor, W3C XML Schema WG <http://www.w3.org/XML/Schema/> 
Member, W3C SML WG <http://www.w3.org/XML/SML/> 
(1-905) 413-3255 T/L 313-3255



"Smith, Virginia (HP Software)" <virginia.smith@hp.com> 
Sent by: public-sml-request@w3.org 

2007-12-05 02:33 AM 

To

"public-sml@w3.org" <public-sml@w3.org> 

cc

 

Subject

[w3c sml] section "resolving an SML reference" section missing.

 

 

 



As promised, attached is a document that contains the missing section in
the SML spec that I believe Kumar was referring to in our last meeting.
The document contains the missing section and also the current text for
comparison. The missing text lists several reference resolution and
model validation requirements. I've indicated in the comments which of
the current sections contain these requirements.

I prefer the current text and organization rather than the missing
section for the following reasons:

- the missing section mixes reference resolution and model validity in
what appears to be a series of 'steps'
- the current organization allows us to define terms and add additional
information along with specifying each requirement.

--
ginny 


#### RefTextComparison.doc has been removed from this note on December
05 2007 by Sandy Gao 
#### RefTextComparison.pdf has been removed from this note on December
05 2007 by Sandy Gao