Robert Sayre wrote:
> it has absolutely nothing to do with interoperation in
> syndication technology. Out it should go.
It *does* have to do with interoperation in that an important
element of interoperation is that diverse processors have a common
understanding of the meaning of the elements that are being passed around.
In this case, it is inevitable that people will do the "intuitively obvious"
think and assume that inheritance works -- even if we say it doesn't. People
will expect it to continue to work if only because they have become
accustomed to it being the expected case with every flavor of RSS as well as
common practice with Atom 0.3.
Thus, even if it makes more "sense" to make the rules that you've
proposed, it will only contribute to interoperability problems in the future
as a result of the diversity of implementation and interpretation that will
result. These issues become particularly important when you look at the
problem of generating composite or aggregate feeds. When you extract an
entry from a feed, what is the correct value for author/copyright that
should be inserted with that entry into the destination feed?
It would make much more sense, I think, to acknowledge that we're
stuck with "inheritance" in at least the ownership fields (author and
copyright) and provide a mechanism to allow people to override the
inheritance when it is inappropriate. This is already assumed to be provided
by having entry-level author/copyright statements trump the feed-level
statements. Perhaps we should complete this system by ensuring that it is
legal to provide null authors and/or copyrights to cover the case where the
feed-level data is known to not apply to the entry yet the "correct" values
are not known. Ie. We could support <author></author>. (The meaning would be
null or "not known" rather than "none".) Or, how does one express "null" in
XML?
bob wyman
