Wil, you ask at the end of your message
below for me to comment on why XDI might specify using XRIs in IRI normal form.
Our original rationale was:
* XDI is based completely on: 1) XRIs for
addressing, and 2) XML for data representation.
* In XML, the anyURI datatype will accept
either a URI or an IRI.
* Since IRI is the least-encoded of the
two forms, we would prefer to use IRI-normal form instead of URI-normal form.
That said, the XDI TC finally decided not to
use the anyURI datatype for the XRI element and instead just specify strings. The
advantage is that this way would be able to specify using XRI-normal form,
which is the least encoded of all.
So the ultimate answer confirms your
suspicion that in most cases the XRI will either be in XRI-normal form or in
URI normal form, and IRI normal form will likely be rare.
=Drummond
From: Tan, William
[mailto:William.Tan@neustar.biz]
Sent: Tuesday, June 06, 2006 10:51
AM
To: Drummond Reed; Sakimura, Nat;
Wachob, Gabe; xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
At yesterday’s call, we discussed my
concern that the application has no way of determining which normal form an XRI
reference was intended to be. That is, even though “@foo%25bar” is
legal in all three (XRI, IRI and URI) normal forms, we cannot tell the “real”
XRI-normal form is because it could either be “@foo%25bar” or
“@foo%BAr” (that underwent XRI-IRI NF transformation).
The problem does not exist between IRI and
URI normal forms because IRI-to-URI normal form transformation is idempotent.
The XRI-to-IRI process, however, is not idempotent but it needed to be so to
overcome the lack of cross references in IRI/URI (I’m sort of guessing
here – someone can clarify).
The conclusion from the call was that, in
general, the XRI parser should leave it up to its caller to specify the normal
form of the input string. The calling application is expected to know the
normal form of its input. For example, if the input came from a HXRI or XRDS
document, they should be in URI normal form. On the other hand, if the source
was user input or some other Unicode-capable XRI library/application, it should
be treated as XRI normal form. I personally think that in practice, there
should be little need for IRI normal form but Drummond mentioned something
about XDI making use of it. Drummond, could you briefly explain?
From: Drummond Reed
[mailto:drummond.reed@cordance.net]
Sent: Tuesday, June 06, 2006 6:47
AM
To: Tan,
William; 'Sakimura, Nat'; 'Wachob, Gabe';
xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
Thanks Les. I will be there, 4pm PT today.
Anyone else who can join in on this issue,
it will be helpful; it’s pretty important from an adoption standpoint.
=Drummond
From: Tan, William [mailto:William.Tan@neustar.biz]
Sent: Monday, June 05, 2006 10:32
AM
To: Sakimura, Nat; Wachob, Gabe;
Drummond Reed; xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
Hi guys,
Les has kindly agreed to open the bridge
for us. Please use the usual conference number:
Dial In
Number: 571-434-5750
Conference ID: 5474
Time:
4pm PT today
Thanks.
From: Sakimura, Nat
[mailto:n-sakimura@nri.co.jp]
Sent: Saturday, June 03, 2006 8:44
AM
To: Wachob, Gabe; Tan, William; Drummond Reed;
xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
I do prefer Monday as well, as I have a
colliding call on your Friday.
Nat
From: Wachob,
Gabe [mailto:gwachob@visa.com]
Sent: Saturday, June 03, 2006 5:40
AM
To: Tan,
William; Drummond Reed; xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
I could do either.
Honestly, I'd prefer 4 on Monday to 4 on
Friday ;-)
Though I have to keep it to about an hour
in either case.
-Gabe
From: Tan, William [mailto:William.Tan@neustar.biz]
Sent: Friday, June 02, 2006 11:45
AM
To: Drummond Reed;
xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
Since no one else has signed up yet, may I
propose Monday 4pm PT?
From: Drummond Reed
[mailto:drummond.reed@cordance.net]
Sent: Friday, June 02, 2006 3:20
PM
To: Tan, William; xri@lists.oasis-open.org
Subject: RE: [xri] Proposal for a
Telecon to close the Normal Forms issue
Wil,
I could do 4pm PT tomorrow (Friday). Can anyone
else join then?
Best,
=Drummond
From: Tan, William
[mailto:William.Tan@neustar.biz]
Sent: Thursday, June 01, 2006 7:13
PM
To: xri@lists.oasis-open.org
Subject: [xri] Proposal for a
Telecon to close the Normal Forms issue
Due to time constraint we were not able to address the issue
that I raised about the XRI/IRI/URI NF issues.
I propose a separate conference call for those who are
interested. Attached is the email on the current thread.
Currently, the only agenda item is:
Does the parser have any issue
distinguishing XRI vs. IRI normal form? If so, how do we address the issue of
ambiguity? Gabe explained that it has been discussed on the TC before and a
conclusion was reached on the matter.
Because I’m in the outback time zone it is difficult
to accommodate both US
east coast and west coast time; the 4pm PT slot seems to work for everyone so
I’d propose that. If possible, we can do one tomorrow (Friday 4pm PT).
Otherwise, any day next week is fine by me too. Another possible time would be
8 – 9am PT.
Thanks.
=wil (http://xri.net/=wil)