W3C home > Mailing lists > Public > public-xpointer-registry@w3.org > January to March 2008

Re: formal issues raised by the smlxpath1 registration

From: Syd Bauman <Syd_Bauman@Brown.edu>
Date: Tue, 4 Mar 2008 20:01:08 -0500
Message-ID: <18381.61652.414189.500969@emt.wwp.brown.edu>
To: public-xpointer-registry@w3.org

MS> (Like Kumar Pandit, I read Syd Bauman's note as a question, not
MS> an objection. Syd, you may wish to set me straight.)

While there is lots about which I may wish to set you straight, this
is not such an issue. Both Kumar and Michael have correctly
interpreted my post as a question, not an objection.


FWIW, I am not completely convinced that outright synonyms should be
registerable idea -- if we have 10 names to mean the exact same
thing, that means that there are 9 fewer names for something else
that might be useful. Overall, however, I agree with Michael that the
bar for registration should not be set very high at all.


KP> To be specific, it only allows xpath1 location paths
KP> to be used. 

OK. That alone seems different enough to me for a newly registered
name. If I had my druthers, the name would be something that
indicates this important difference. E.g., I would have named it
xpath1location() or locationpath1() or xlocpath1() or some such.


KP> The SML working group decided to define smlxpath1 scheme
KP> for the following main reasons: 

KP> 1. The SML group did not want the SML specification to take
KP>    normative dependence on xpointer schemes that are not W3C
KP>    recommendations. The xpath1() scheme is not a W3C
KP>    recommendation.

Um ... is smlxpath1() going to be a W3C recommendation?


KP> 2. The xpointer schemes that is a W3C recommendations
KP>    (namely, element()) does not meet the needs of SML.

KP> 3. The xpath1() scheme allows any xpath expression. The SML
KP>    working group did not want that. The group wanted only xpath
KP>    location paths. 

KP> 4. The smlxpath1() scheme inherits namespace bindings from the
KP>    containing element which greatly reduces verbosity of SML models
KP>    because xmlns() portions do not need to be added to each SML
KP>    reference.

Oh. That's nice. But in some sense it isn't XPath1, then, right? That
is, I thought

       It is an error if the QName has a prefix for which there is no
       namespace declaration in the expression context.
                        -- XPath 1.0, sect 2.3

meant that you're not allowed to do this. I am not at all sure I'm
interpreting the spec correctly -- that the namespace prefix must be
declared within the same XPath expression -- and if I am, I know I
have no idea why that restriction exists.
Received on Wednesday, 5 March 2008 01:01:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 5 March 2008 01:01:28 GMT