Re: URNs approved by Karl

[John Cowan <jcowan@reutershealth.com>]

Norman Walsh wrote:


> Versioning is open issue #26. The namespace name is one way to do it, but
> it's fairly coarse. It means that 1.0 processors will be completely
> unable to recognize 1.1 catalog documents.

That turns out not to be the case, as Kevin Renner says.

If an element remains unchanged from 1.0 to 1.1 in both syntax and
semantics, there is no reason to change the prefix on that element.

> I think I prefer a required
> version attribute on the catalog element. Then we can say that version 1.0
> processors are allowed to ignore elements that they don't recognize when
> processing version 1.0+delta catalogs.

But what if the semantics of some element changes?  In that case, the
1.0 processor would *misinterpret* the 1.1 catalog.  This would be
far worse IMHO than barfing on it.

(Of course, there will never be a 1.1, as we all know.  :-))

-- 
There is / one art             || John Cowan <jcowan@reutershealth.com>
no more / no less              || http://www.reutershealth.com
to do / all things             || http://www.ccil.org/~cowan
with art- / lessness           \\ -- Piet Hein