On Wed, May 7, 2008 at 5:20 PM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
> > But if I am outputting a text document with iso-8859-1 then
>  > the presence of non-iso characters in the output will raise an error.
>
>  The spec says "should" rather than "will". XSLT 2.0 is stricter, it says
>  "must".
>
>  If you're asking what the spec says, that's the answer.
Hmm, yeah, are there any processors that don't implement the Should as
a requirement?

> If you're asking why
>  it says that, that's a different question.
no, I was basically asking if my interpretation was approximately
correct, or what the issues in play were. so  that's cool.
> It's very hard to answer
>  questions about why a WG made a particular decision even if you were at all
>  the relevant meetings - which I wasn't. What one can do is try to invent a
>  rationale that the WG might use to justify the decision if asked.

yeah, the same thing I was thinking. I could argue it either way actually.


One thing I was wondering about, since it says MUST in 2.0 is there an
argument there?

Cheers,
Bryan Rasmussen