How about a [EMAIL PROTECTED]"homepage"] for feeds that actually have a home page to point to. By the same reasoning, [EMAIL PROTECTED]"alternate"] would exist if the feed actually had an alternate representation. Both of these would be MAY. This would be consistent with how the spec currently treats the prev, next, etc. with the verbage "atom:feed elements MAY contain additional atom:link elements beyond those described above" Sec 4.1.1.
If all of these links are MAY, then the only effect of the client is the ability to have more features if they are included. I can't think of an example where the client would fail to function if one if them would be excluded (no reason for a MUST).
Antone Roundy wrote:
On Sunday, April 3, 2005, at 05:16 PM, Robert Sayre wrote:
Tim Bray wrote:
Well, yeah, but when they do the half-hour's coding it's going to cost them to start supporting Real IETF Atom 1.00 (tm), they can do an extra 3 minutes and if there's no <link>, they don't make the subscription clickable.
I doubt they've never encountered a feed without a link. Why do you think they didn't fix it?
Perhaps a combination of:
* Because the RSS spec says its required
* Because it doesn't cause any serious problems like a computer or application crash, just an unexpected page to be shown when the link is clicked
* Because nobody who we assume pointed it out to them complained loudly since it didn't cause any serious problems
* Because they had more important things to do
Both the amount of work required to enable an app to handle linkless feeds and the negative impact of not having the link are trivial as far as I can see, except for Bill's point about wanting an identifier for the feed.
On Monday, April 4, 2005, at 06:25 AM, Bill de h�ra wrote:
Anyway I've made my position clear at this point. Please make id or self mandatory.
I'd be opposed to [EMAIL PROTECTED]"self"] or [EMAIL PROTECTED]"alternate"] being mandatory since uses cases have been put forth for not having a reasonable value for either, but not opposed to either or both being SHOULDs. I also wouldn't be opposed to a feed being required to have either a [EMAIL PROTECTED]"self"] or an id.
