[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xri] RE: #next: alternate to #replace
Eran Hammer-Lahav wrote on 2009-09-18: > This thread is becoming impossible to follow. The main reason for that would be HTML email, which I beg all of you to stop, cease, desist... On topic, I'm trying to follow/understand things and have a few short notes: I think the delegation concept is important. I see it as likely for people to want to override portions of a generic XRD. It's also fine if that's a custom link type left out of this spec. But I don't understand this idea that Links don't result in something with Type information you would process. Following a link to an XRD should result in a new XRD that gets processed according to the semantics of XRD, which would include Type selection. I think XRD processing ought to be document/instance-atomic and that trying to *prevent* differences in the result when you encounter a set of XRDs in multiple orders is impractical. Order will always matter. So in light of those thoughts, if there's some use case that isn't met, I would analyze it in terms of how to solve it without violating those rules. My limited input, -- Scott
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]