I do think also that there are plenty of use cases suggesting that alternate might not be always available. However in that context where there's no feasible alternate, it would be extremely useful to know you can count on atom:id. If not, what else can you rely on to track your "content" as it flows from system to system without an alternate. I have not been following atom-syntax for a while, but are there any other elements that could be used if no alternate was not found? I think this is the reason why Sam has never seen anybody complain about link because, it doubles as id as well, but since we have a clear separation in atom, either or both might be good to be specified.
Just my two cents. Elias Torres On Apr 3, 2005 7:04 PM, Bill de hÓra <[EMAIL PROTECTED]> wrote: > > Tim Bray wrote: > > > > On Apr 3, 2005, at 12:45 PM, Graham wrote: > > > >> So do you have an argument here as to why it should be required? All > >> I'm seeing is that it's easy to workaround when the publisher omits it. > > > > > > Agreed. Joe, that wasn't very convincing. I repeat, we've seen several > > very believable use-cases for why someone might want this, and no good > > arguments (that I can remember) that it would break anything. Sam has > > pointed out that no previous version of RSS has done this, which is a > > reasonable argument; except for we have use-cases, and nobody's shown > > that the cost is non-zero. -Tim > > The top level feed stuff doesn't make a lot of sense to me. That > @alternate is mandatory while @self and atom:id are not isn't very > convincing either: > > atom:[EMAIL PROTECTED]'alternate'] : MUST > atom:[EMAIL PROTECTED]'self'] : SHOULD > atom:id : MAY > > You can't have a useful discussion about @alternate without talking > about @self or atom:id. > > Thus: I'm -1 to downgrading @alternate unless @self is lifted to MUST or > atom:id is lifted to MUST. If either are lifted to must I'm 0 on > downgrading @alternate. At that stage @alternate doesn't matter a whole lot. > > cheers > Bill > >