Andreas Sewe wrote:
While I do realize that the name "Slug" is quite entrenched in a publishing context, I still wonder whether a more neutral name like "Segment" would be better? IMHO, it is conceivable that other protocols might want to reuse the header, in which case a name like "Segment" is far more appropriate than "Slug". Any opinions?
It's not just monikering, processing segments is more work. If you're serious about passing IRI/URL/URL/WhoKnows fragments through this header, as opposed to a string that needs to be downcoded to ascii, which is what all the deployed slug code (I've seen) does, I'd like to see done done outside the core protocol. Then we can have Segment: torture tests in 2007. Please decide soon tho' :) I have an action to write the security notes on this feature.
cheers Bill
