W3C home > Mailing lists > Public > public-qt-comments@w3.org > August 2002

xf:compare collationLiteral is an anyURI?

From: Joseph Kesselman <keshlam@us.ibm.com>
Date: Wed, 21 Aug 2002 14:27:02 -0400 (EDT)
To: public-qt-comments@w3.org
Message-ID: <OF6BA86E7C.51220599-ON85256C1B.004FEFFE-85256C1C.0065D14E@us.ibm.com>


The spec says xf:compare collationLiteral is an anyURI. But your examples
      xf:compare('Strasse', 'Straße', anyURI('deutsch'))

If that's really a URI, it's a relative URI reference and needs to be
interpreted relative to a base URI. But this example strongly suggests that
you intended it to be taken only as a literal string.

I submit that if it's a literal it shouldn't be called a URI, to avoid this
confusion. Remember the massive debate over whether relative URIs were
meaningful as namespaces? You _really_ don't want to go there.

Conversely, if it really us a URI, you shouldn't be using the relative form
in this example since (a) it's misleading to the reader and (b) it's not
typical use -- you generally wouldn't want a document's collation to change
if the document is moved to a new base URI.

Joe Kesselman  / IBM Research
Received on Wednesday, 21 August 2002 14:54:04 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:56:43 UTC