Unique IDs allow clients to determine the state of the feed. If entry ids are
not unique, then we still need some other way to determine the unique
state of the feed. If we allow duplicate IDs but *require* something else
to be different (e.g. update time), then we can still determine the unique
state of a feed and repeated IDs are OK. We would need to properly
document which elements make an entry unique in the event of a duplicated
ID. Brett.
On 6 May 2005, at 2:10 pm, Dave Johnson wrote:
Yes, I think both of my arguments fail to hold and I no longer have a real objection to duplicates. Allowing duplicates gives feed produces to model events or other objects (versioned documents in a wiki) as they wish. Like you, I wonder "Does anyone remember why having the same id in a feed is a bad idea?"
