Tim Bray wrote:
On May 3, 2006, at 3:55 PM, Thomas Broyer wrote:
I also still don't understand why the response body, if provided,
"must" be an Atom Entry Document:
Bah, we discussed this a *lot* and I thought it was pretty clear the WG
had consensus for an Atom entry or nothing. -Tim
I read what Mark Baker had to say about this and I'm inclined to agree
with him. It's short-sighted to assume Atom is all anyone will ever need.
On the other hand one reason to insist on Atom is that currently the
data format is fundamental to the workings of the protocol machine, cf
any required Atom field has to be supported. The media paces indicate we
are not doing that properly yet, as did past discussion around atom:id.
(I'm not beyond persuading that this dependency a design flaw)
Thus - I'm fine with SHOULD, *iff* someone can explain to me how a
SHOULD directive works sensibly with headers like Title or elements like
atom:id.
cheers
Bill
