Feedback on WD03

["Gabe Wachob" <gabe.wachob@amsoft.net>]

 

Here’s my notes – they are against the word doc with the edits/changes marked

 

Line 27 (table 1): I think we need to define “local” now since its use has been deprecated?

Line 30: I think we need to define “local” now since its use has been deprecated?

 

Line 62: s/web servers/web infrastructure/

 

Line 384: define “target resource”

 

Line 389-392: Instead of using the MUST language, isn’t it more comprehensible to say “is relative to the authority producing this XRD”?

 

Line 394-395, 399-400, 404-405: the language “MUST be an absolute identifier MUST be assigned by the same parent authority” doesn’t seem to make sense – segments get assigned, not absolute identifiers…

 

Line 413: the word “management” in the title of section 4.2.4 seems completely redundant/useless? These are simply Service Endpoint Elements, right? How about Service Endpoint Descriptors?

 

Line 430, 445, 465: I’d make these sections children of 4.2.4 rather than peer following sections

 

Line 568 vs 572: I might be reading this wrong, but it seems inconsistent that  in line 568, you can use either, but under line 572, you MUST begin with HTTPS HEAD?

 

Line 571: What “protocol” are we talking about here?

 

Line 1696: (Section 11) From a document flow point of view, it seems like this section should be right after section 4 (XRDS documents)

 

Line 1728: Figure 8 – the lines from a node to another line are not really well defined – the thing is that the line (“forward reference”) is not pointing to a single edge, but to an entire path of edges… not sure how to demonstrate this in the graph but the current graph is a bit misleading…

 

Line 1757: s/authority/resource/ ?

 

Line 1765: What does “edit” mean here?

 

Line 1786: What does “edit” mean here?

 

Line 1896: Can two XRD’s Ref each other instead of having a directed relationship?

 

Line 2105: Section 13 seems like it should go before sec 12?

 

Line 2113-: this is odd – the parameter is only for proxy & recursive resolvers, right? This wouldn’t apply everywhere would it?

 

Line 2213: Define what a “service reference” is?

 

Line 2225- 2231: Why are only HTTP URLs allowed here?

 

Line 2427-2429: I think I’ve raised this before, but adding this rule makes verifying XRDS results more complicated and seems like an ugly hack