Re: [svg2] request for review of new SVG properties

On Sep 13, 2012, at 9:42 AM, Simon Sapin <> wrote:

> Le 13/09/2012 08:30, Tab Atkins Jr. a écrit :
>> On Wed, Sep 12, 2012 at 10:25 PM, Sebastian Zartner
>> <> wrote:
>>> Dirk is right. The CSS specs should be changed to clearly specify that
>>> IRIs are allowed. There were already other discussions about this
>>> [1][2].
>>> A quick test (using one in "background-image") showed that all major
>>> browsers (IE, Firefox, Chrome, Opera, Safari) already support IRIs. So
>>> I suggest you replace the <url> definition by <iri> throughout the
>>> specs and give a clear definition of what it implies.
>> We're not going to change the name, because <iri> is a stupid name
>> that nobody outside some rarified standards circles ever uses.  The
>> thing that you use for links is called a URL in common and most
>> technical parlance.
>> I have no problem with clarifying the definition if necessary,
>> particular if it's merely to reflect current implementations.
> I think that Dirk was not asking to rename <uri> to <iri>, but only 
> clarify the definition and make <uri> accept IRIs, not just ASCII URIs. 
> With that, it becomes the same as <funciri> SVG and they can be unified.
That is correct. I am fine with replacing <funciri> with <url> and just want to state that the definition of <url> should be updated to support IRI in general.

I won't be able to attend the next meeting. I would like to have it clarified in one place in the long term (CSS values and units?). For the moment I would be fine if the CSS WG can resolve that all affected specifications add a short sentence with clarification in the "Values" section (currently Filter Effects, CSS Masking, SVG and possibly CSS Exclusions).


> (The separate issue remains of deciding what unicode strings exactly CSS 
> considers a "valid IRI/URI" and what to do with the other ones, but 
> that’s for another thread, as Sebastian noted.)
> -- 
> Simon Sapin

Received on Thursday, 13 September 2012 09:41:03 UTC