Re: [Bug 5164] validation vs assessment

I haven't had time to read the proposal itself, and may not until at least 
after TPAC, but FWIW the gloss you provide below seems fine to me.

> One aspect of the proposal that will need discussion is the 
> choice between a
> single definition of 'valid document' and a set of several 
> definitions which
> capture different conditions

I don't think I have a strong feeling on this.  Anything your or the WG 
comes up with is likely to be fine with me.

What does occur to me is that the same community that will value having 
this terminology also tends to be confused about the distinction between 
"schema" and "schema document".  I don't have any great ideas for how to 
help.  Noodling on the question, I wonder whether it would on balance be 
helpful to remind them with a note along the lines of:

----
Note:  as with other definitions of validity in this Recommendation, "root 
validity" of an instance document is determined with respect to a 
*schema*, not a a *schema document*.  Thus, it is not in general 
appropriate to speak (hmm, not the best word) of instance document I being 
"root valid" with respect to schema document sch.xsd;  it is typically 
appropriate to speak of instance I being "root valid" against the schema 
that results from the composition (right word?) of schema document 
sch.xsd, or perhaps schema documents sch1.xsd, sch2.xsd, etc.   with the 
built in schema components.  If sch.xsd contains <imports> and/or 
<includes> with schemaLocation attributes, then it is generally necessary 
to indicate which if any of the referent schema documents are to be 
(transitively) composed to create the schema.
---

I'm not at all happy with the wordiness or clumsiness of the note proposed 
above;  I'm just pointing out some regret that defining "root validity" 
will help, but I bet we'll still see a lot of specs saying "your instance 
must be root valid against a.xsd".  I wish I had a better proposal for how 
to help.

Noah


--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








bugzilla@wiggum.w3.org
Sent by: www-xml-schema-comments-request@w3.org
10/29/2009 02:09 PM
 
        To:     www-xml-schema-comments@w3.org
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        [Bug 5164] validation vs assessment


http://www.w3.org/Bugs/Public/show_bug.cgi?id=5164


C. M. Sperberg-McQueen <cmsmcq@blackmesatech.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|needsPublication            |needsReview




--- Comment #10 from C. M. Sperberg-McQueen <cmsmcq@blackmesatech.com> 
2009-10-29 18:09:52 ---
A wording proposal intended to resolve bug 5164 and bug 6015 is now on the
server at

  http://www.w3.org/XML/Group/2004/06/xmlschema-1/structures.b5164.html
  (member-only link)

Its salient features are:

1 Drops the claim that the spec uses 'valid' only to refer to local 
validity;
this is just not true.

2 Adds a definition of validation that makes clearer the intimate relation
between 'validation' and 'assessment'.

3 Points out that in all cases, validation and assessment overlap somewhat
operationally, and in many cases the two terms are extensionally 
equivalent. 
Says that when we wish to emphasize the calculation of the [validity] 
property,
we use v-words, and when we wish to emphasize the rich fullness of the 
PSVI, we
use a-words.  And when no particular emphasis is intended one way or the 
other,
the choice is arbitrary and often based on historical accident.  (E.g. the
v-word not an a-word in "PSVI".)

4 Changes a few uses of 'validation' to 'assessment' to help make 3 true.

5 Changes all the cases where the spec currently says "X must be valid 
against
Y, as defined in Rhubarb Locally Valid (section 89.23.2)" to say that X 
must be
*locally* valid, since local validity is what the rule referred to 
actually
defines.  These cases (ten or twenty of them) are the cases where the 
current
text actually does use 'valid' to refer to local validity.

6 Provide a definition of 'valid document' (OR: of root-valid document,
deep-valid document, and uniformly valid document) for the convenience of
working groups who use XSD to define XML vocabularies and want to say "to
conform to our spec, your document must be (root|deep|uniformly)? valid 
against
our schema", instead of having to specify the initial conditions for 
validation
and the required results of validation using terminology they often feel
uncomfortable with.

In addition, some editorial changes were made. 

One aspect of the proposal that will need discussion is the choice between 
a
single definition of 'valid document' and a set of several definitions 
which
capture different conditions (root element has [validity] = valid, vs. 
root
element has [validity] = valid and no node in the document has [validity] 
=
invalid, vs. every node in the document has [validity]=valid).  And if the
latter, the choice of terms must also be considered. 


-- 
Configure bugmail: http://www.w3.org/Bugs/Public/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.

Received on Thursday, 29 October 2009 19:43:30 UTC