Re: Towards a consensus draft (urgent)

On 9/11/07, Norman Walsh <ndw@nwalsh.com> wrote:
> | In order to move things along, what would make me "comfortable"
> | would be that we have the general language that Norm has
> | put forth and that we require *our* steps to output "infosets"
> | that don't require namespace fixup.
>
> I believe the WG has rejected that position.
>
> | I'd be OK with leaving that as "implementation defined" but we could
> | easily have a non-normative appendix suggesting ways in which
> | you might go about it.  Since such text would be non-normative, it doesn't
> | have to be 100% correct.  I think we can get very close with the proposal
> | I sent out earlier and I'd be fine with that being a non-normative suggestion.
>
> Do I understand correctly that you're saying that you'd be happy if we
> made the detailed suggestions about how to achieve output that doesn't
> require namespace fixup non-normative and left the question of whether
> or not implmentations produced infosets that require fixup
> implementation-defined?
>

Well, I wanted the combination our steps being required to do the right
thing.  I don't quite understand why people believe this is too much of
a burden in that if a serialization engine can do namespace fixup, surely
a step can do it much easier as their changes are local.  The performance
burden is a big red herring as we've rejected that argument for other
parts of our specification and added features which require a lot more
processing than adding a few simple namespace declarations.

...but "You can always get what you want, ..."

I settle for non-normative guidance in an appendix.

-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

Received on Tuesday, 11 September 2007 15:56:01 UTC