Re: When and how are xs:anyURI references made absolute?

Hello Norm,

Thanks, that makes sense, or enough sense. :-)

IMV the root of the problem is that xs:anyURI has underspecified semantics.
There is no rule for how a relative URI should work. That is part of its
charm, I guess.

In general, I am in favor of (a) fulfilling users' expectations as a
preferred option, and also (b) correcting broken semantics wherever
possible in favor of improvement and viability, sometimes even at the cost
of strict backward compatibility. These aims can be at odds.

If the whole situation can be improved by tweaking the definition of a rare
(but essential) step such as namespace-rename, I am even okay with that -
if the improvement is clear and impacts are really limited. (For now.
Changing step semantics is a bad idea going forward.) OTOH, any solution
that requires explanation to users (of how to do something or why something
is different) would make me sad.

I also think that warnings at least in verbose or trace mode might play a
role here, in mitigating impacts if any.

Not sure that's helpful --
Wendell


On Fri, Jun 27, 2025 at 2:50 PM Norm Tovey-Walsh <ndw@nwalsh.com> wrote:

> Wendell Piez <wapiez@wendellpiez.com> writes:
> > Has this issue been there all along (in XProc 1.0), or if not, what has
> changed?
>
> XProc 1.0 didn’t allow you to specify the types of options, so the
> question can’t arise in that context.
>
>                                         Be seeing you,
>                                           norm
>
> --
> Norm Tovey-Walsh <ndw@nwalsh.com>
> https://norm.tovey-walsh.com/
>
> > Not everyone can live upstream.
>


-- 
...Wendell Piez... ...wendell -at- nist -dot- gov...
...wendellpiez.com... ...pellucidliterature.org... ...pausepress.org...
...github.com/wendellpiez... ...gitlab.coko.foundation/wendell...