OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Requirements Feedback from Visa


Drummond-
	Here's a summary of feedback from folks at Visa. This is summarized
to make the work of integrating this stuff easier. Most comments were in the
Motivations section. 

GENERAL COMMENTS

* Do we need to mention addresses that support the concept of "multicast" or
"broadcast"? More generally, do we need to have a concept of "groups" in XRI
(as distinct from individual resources). 

LINE-BY-LINE COMMENTS

117: The Motivations section needs to flow really smoothly and be really
tight. Needs some more work.

119: Explain the "100% concrete/100% abstract" distinction. Is this really
true? Is "abstractness" really the distinction? Or is it more
"logical/physical" a distinction? 

120+: In introducing the layers of Figure 1, introduce each in a bullet
point to make the paragraph clearer and flow better. Also, make sure that
"addressing" is described consistently and tied in with each layer (e.g.
what is each layer *addressing*).

152-180:  Its confusing to talk about abstract identifiers but also mean
persistent and reliable too. How is abstractness equivalent (or is it?) to
persistent and reliable? 

163: The urn given is a bad example of a identifier which is non-semantic
and hard for humans to use. In fact, I think it *is* semantic and not *too*
hard for humans. Most URIs in use today are semantic and easy for human use.
A good example that's both non-semantic and hard are the HTTP URLs for some
sites which have complicated URLs -- e.g.
http://www.business2.com/articles/web/0,1653,48716,00.html

172: Figure 2- The top two elements should perhaps be called "Abstract
Identifiers" (maybe also on figure 1)

182+: A even-better example is needed - perhaps this section focuses too
much on trying to describe a physical analogue to cross-context
identification. Maybe talk about the "same invoice in different enterprise
systems" example?

201: "concatenating" is too narrow and/or misleading - "concantenation" is
just a mechanism or a "means to a end" - describe the more general process
going on here of syntactic inclusion. 

212: We need to be careful about the term "resource" - probably should talk
about resource data here (what does "instances of resources" mean?) 

218: change "would" -> "might" 

255: "to contain no data..." is awkward. need to emphasize that the "free of
personally identify information"

313: This first sentence uses "identi*" too much - can easily lose the
reader. 

367+: Requirement here is really "IDs must support the represetnation of
objects in the users's native language and scripts" - what is here is an
implementation (though the only obvious one) of that requirementrs

471+: This is perhaps and awkward placesment of this requirement (GMW: I
don't remember who's comment this is, but I can't figure out a better
placement)

476+: Probably should lose the word "veronymity" in this requirement since
a) nobody knows about it and b) it should probably be stated as (at the end
of the rquirement) ".. and support the ability to make assertions about the
verfiability of the presented identifiers" (or something like that).
However, isn't this outside the scope of the XRI spec? 



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]