Re: p:namespaces proposals

/ Jeni Tennison <jeni@jenitennison.com> was heard to say:
| Norm,
|
| On 22 Jul 2008, at 02:15, Norman Walsh wrote:
|> Currently, we say it's a static error (err:XS0051) if the value of the
|> exclude-prefixes attribute is wrong. I've changed that into a dynamic
|> error because you can't check the prefixes until runtime.
|
| Huh? I assume you're talking about this:

Sigh. Nevermind. Writing code on a plane: yes; editing a spec on a
plane: not so much.

|> We also say:
|>
|>  It is a dynamic error (err:XD0013) if the specified
|>  namespace bindings are inconsistent; that is, if the same prefix is
|>  bound to two different namespace names.
|>
|> Do we really want that? Does it make more sense to construct the
|> mapping in the order in which the p:namespace elements are specified
|> and just say that if the same prefix is defined multiple times, the
|> last binding wins?
|
| I'd be OK with that as long as there's wording to imply that the
| processor should give the user a warning in that case.

I don't think we can tell implementations that they must issue warnings.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Be indiscrete. Do it continuously.
http://nwalsh.com/            |