POWDER 2.1.3 review

Grosso, Paul scripsit:

> ACTION to John:  Review section 2.1.3 of the POWDER spec at
> http://www.w3.org/2007/powder/Group/powder-grouping/20090309.html#canon

I've looked at it.  This section describes an algorithm for translating
valid IRIs into other valid IRIs; for example, it downcases scheme and
host names, decodes useless %-escapes, removes redundant port numbers,
and so on.  It says nothing at all about strings that are not IRIs.  If
applied to a LEIRI that is not an IRI, such as "http://example.com/5>3",
the result would be "http://example.com/5>3", which is also not an IRI.

Therefore, I conclude That it is independent of the LEIRI spec and that
the LEIRI spec is not relevant here.

This email discharges my formal action.

HOWEVER, there is another problem.  One of the canonicalization rules
of the algorithm says "+ characters in the query string are converted
to spaces".  This rule transforms a valid IRI into an invalid one, as
valid IRIs may not contain spaces.  I propose that we suggest to the
POWDER WG that they reverse the rule to read "Space characters in the
query string are converted to + characters" instead.

I ask that this point be added to the agenda.

-- 
Using RELAX NG compact syntax to        John Cowan <cowan@ccil.org>
develop schemas is one of the simple    http://www.ccil.org/~cowan
pleasures in life....
        --Jeni Tennison                 <cowan@ccil.org>

Received on Wednesday, 8 April 2009 16:32:22 UTC