Antone Roundy wrote:
On Friday, February 4, 2005, at 03:04 PM, Robert Sayre wrote:
Sophisticated implementations can assert common ancestry with an extension element.
Could you elaborate a little on this. If we forbid putting two revisions of the same entry into a single document (unless you change the ID between revisions...please, I hope that's not what's being suggested), then what could an extension do?
It's a trade-off. Make atom:id confusing to implementors and have poor interop with simple code, or require more complex code to jump through hoops. So, my suggestion is exactly what you'd feared. You could do an end run on the problem and pick a different delivery mechanism.
Robert Sayre
