Re: Comments on XML Infoset 20 December 2000 WD

Thank you for your very useful comments on the Infoset draft.

> The nature of stringness, or, The synonymy of nullity and emptiness?

As you note, we don't use "not present" any more.  Everything that can
sensibly not have a value says (or should say) that it is null in that
case.  I'm not sure how we should define null - "a value distinct from
all others" perhaps?  We should certainly note that it is different
from the empty string, which is just a string with no characters.
I don't think we need to define "string" do we?

> Pre-validation infoset state

We already know that there are a number of issues concerning
non-validating parsers, especially ones that don't read the whole DTD.
We haven't decided what to do about them yet.

> Inconsistency between (2.3 and 2.15) and (2.6, 2.13 and 2.14)

We removed the [children] property from attributes, and now use
attribute IIs to represent namespace declarations (as opposed to
in-scope namespaces).  So these inconsistencies are just stale
references that should be removed.

> Political correctitudinality, gender-wise

Easily fixed.

> CDATA sections

We will take note of your comments about round-tripability.  I don't
understand your comment about redeclaring predefined entities - at
least as of XML 1.0 second edition the allowed declarations are very
limited.

> Finally, some typos

There's been a lot of global-replace-regexp lately...

Thanks again,
-- Richard

Received on Wednesday, 3 January 2001 09:49:58 UTC