RE: [xri] Draft -07 feedback from another Visa person

["Wachob, Gabe" <gwachob@visa.com>]

[3rd time's a charm?]

 

Initial responses to comments below:

---------

XRI Comments

 

340-341 Why are cross references only absolute?  For simplicity or
        conciseness?  Is this just a nomenclature issue?
        i.e. identifiers (relative, etc) satisfy local
        cross-referencing needs? 

Cross references are absolute mostly because its not clear what the context for the identifier would be if it *weren't* absolute. That is, if a relative identifier were used, then what root namespace would have been defined in? 

 

I think there may also be some logical reasons having to do with the motivations for cross references. That is, the purpose of a cross reference is to use an identifier defined elsewhere in another namespace/identifier community. Coming from "elsewhere" implies, generally, that an identifier needs to be globally uniquely identified -- thus the requirement of being absolute.

 

 

146     Is scheme interoperability with HTTP a requirement or fringe?

My personal opinion is that its not a core feature or requirement, though it could be important in the transition to using XRIs.

 
482-483 Scheme also indicates a connection/transport protocol.  In XRI
        this is lost, which eliminates the possibility of inferring a
        default port.  I would argue that you might want to add some
        language here like:
 "Within a federated namespace, the omission of a port may
 indicate a default port.  This use is at the discretion of
 the governing namespace authority (and implementing applications)." 

 

Well, a host name is not simply neccesarily resolved into an A record and then connected as a simple TCP connection. That is, in the DNS-specified naming authority resolution, we use SRV records, which include port numbers. Thus, this language about being "undefined" should probably say "meaningless" or "irrelevant" as well as undefined. Port numbers are specified in the resulting SRV record.

 

511 Is xri a valid URI scheme for an xref? 

Any URI scheme is valid. If you choose to embed an XRI URI as a cross reference, then it is treated with the same escaping rules as other URIs and appears basically as an opaque URI to XRI processors. Normally though, you'd insert just an XRI value (without the scheme) which makes for a more readable, compact syntax for cross references. For example:

 

xri:/1.2.3/(=Gabe)  

 

vs.

 

xri:/1.2.3/(xri:=Gabe)

 

 

519 Why define global-xri here? I don't see the relevance. 

see the answer to the previous question 

 

521 First use of "node"?  What does it mean in the context of XRI? 

terminology out of sync ;-) I believe the terminology this corresponds to in the BNF is "sub-segment".  

 

525 Is this to imply that only when the entire XRI is in () that
 it is unresolvable?  I would presume that XRIs can include
 cross-references in path-segments that are also unresolvable. 

This sentence only implies that the following syntax indicates the entire XRI is unresolvable:

 

xri:(/foo.bar/baz)

 

As for cross-references being "unresolvable" - this is really a misinterpretation of "unresolvable". Unresolvable means only that the person creating or writing down the XRI doesn't intend for anyone to actually resolve the XRI. Whether or not the XRI can be put through the resolution process or not is an existential question that depends on deployment (ie are there directory entries for each sub-segment?).

 

Cross references are never intended to be interpreted in any way other than as an opaque string. Thus, a "non-resolvable cross reference" is a redundant statement.

 

I think the problem is the term "non-resolvable" - it should be "not-to-be-resolved"..

 

550-551 I didn't get the meaning of persistent or re-assignable identifiers
 out of this description.

 

2.2.3.1 Is it allowable to escape unicode characters?  For example, if one
 wanted to express an international XRI in IA5 (ASCII)?  In this
 case, the %AB format described in 2.2.3.1 is insufficient to support
 the expanded character width.

 

694 Does the lack od idempotency affect semantics or syntax?  I would
 hope it would only be syntax.

 

2.2.3.3 How about this as an alternative?
     Escape all current escapes (%s).
     Escape all syntactic elements with cross references (parens).
     Escape all parens.

 

878-879 Why are XRI authorities compared in a case-insensitive manner?

 

Section 3   (I still need to do some reading)

 

Global:

 

 Has there been any work on DECODING XRIs?  It's not immediately
 clear from the ABNF that decoding is unambiguous.

 

 In addition, aside from unresolvable references, is it possible
 to canonicalize XRIs?  This is a highly desireable feature
 (for equivalence, at a minimum).

 

 An XRI is not a URI (because of the expanded syntax).  But
 is an URI an XRI?  (no, because of different scheme (xri)).
 I think it would be nice to all URIs be valid XRIs.