RE: [office-formula] Summary 2010-09-07 - IRI vs URI

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

There's a bug being perpetuated here. 

1. Every IRI is a URI

2. Not every URI is an IRI.

The IRIs are subsets of URIs, not the reverse.  

What IRI does is introduce a specific %-escaping principle by which Unicode
code points that do not have URI-allowed ASCII code points can be converted
to UTF8 and then any individual UTF8 bytes that remain non-URI-allowed are
%-encoded.  This is very different from other uses of %-encoding that is
also allowed in general. In particular, not every %-encoding has a decoding
as valid UTF-8-encoded Unicode.  The cited press-release statement is bogus
for our purposes.  

It is also important to recognize that an equivalent of this rule was
adopted for XML Schema datatype "anyURI" (and/or, I believe, the xhtml:href
value) before the IRI RFC was ratified.

With regard to Zip, using IRI encoding is the safest possible thing to do,
since it is a single-byte character encoding that stays in a very safe ASCII
subset, something that should work in Zip everywhere.  I recommend it.

-----Original Message-----
From: David A. Wheeler [mailto:dwheeler@dwheeler.com] 
Sent: Tuesday, September 07, 2010 09:01
To: office-formula@lists.oasis-open.org
Subject: [office-formula] Summary 2010-09-07 of OpenFormula meeting

Summary 2010-09-07 of OpenFormula meeting

[ ... ]

* OFFICE-3342 ("IRI")
http://tools.oasis-open.org/issues/browse/OFFICE-3342

The Japanese are concerned about this;
clearly we want to make sure that these specifications
properly handle international issues like this.

Wheeler:
http://www.w3.org/2004/11/uri-iri-pressrelease.html.en
says that "every URI is already an IRI."
Therefore, IRIs are a superset.
Therefore, we should switch everything to IRI for maximum flexibility
where we can.

Weir: Perhaps not in the zip packaging; might not be able to handle IRI.

Wheeler: Fair enough.  But for formulas, we should accept IRIs not just
URIs.

Patrick: We need to insert a normative reference for IRIs.
IETF RFC 3987, and that is still current.

Wheeler: Patrick - please add a comment to switch in part 2 all URIs/URLs to
IRIs.

[ ... ]