W3C home > Mailing lists > Public > www-style@w3.org > August 2012

Re: [css3-exclusions] Use <funcIRI> instead of <uri> on path referencing

From: Dirk Schulze <dschulze@adobe.com>
Date: Tue, 28 Aug 2012 15:39:27 -0700
To: "Tab Atkins Jr." <jackalmage@gmail.com>
CC: www-style list <www-style@w3.org>
Message-ID: <A97BBF79-7BE7-44FB-ABDF-0EDE80EF3B0D@adobe.com>

On Aug 28, 2012, at 12:43 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:

> On Tue, Aug 28, 2012 at 11:50 AM, Dirk Schulze <dschulze@adobe.com> wrote:
>> On Aug 28, 2012, at 8:52 AM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
>>> On Tue, Aug 28, 2012 at 6:50 AM, Dirk Schulze <dschulze@adobe.com> wrote:
>>>> Hi,
>>>> 
>>>> The CSS Exclusions specification uses <uri> to reference SVG shapes[1]. This should be changed to <funcIRI>, to be consistent with SVG and Filter Effects.
>>> 
>>> It should actually be using <url>, as that's the name defined in CSS3
>>> Values (we've been inconsistent in the past about which name to use).
>>> 
>>> SVG's <funcIRI> is identical to CSS's <url>, except more confusingly named. ^_^
>> No, <url> takes an <URI> as input [1], <FuncIRI> takes an <IRI>[2] as input. <URI> is a subset of <IRI> and SVG (even in the current WD) requires IRI support. Even so they can be converted to each other. Furthermore SVG1.2T has the following sentence: "Other W3C specifications are expected to be revised over time to remove these duplicate descriptions of the escaping procedure and to refer to IRI directly."
> 
> The differences between a URL, URI, and IRI as defined by the various
> standards bodies are irrelevant to authors.  CSS doesn't care - it
> takes whatever browsers choose to accept as a "url".  The grammar of
> url() is basically just "url(, then anything that's not a )
> character", which encompasses all of the various silly terms.  We just
> use the "<url>" grammar token for it, because that's the term most
> familiar to authors.
According to CSS 2.1[1] url() on CSS uses URI, as specified by [2]. Which is just a subset of IRI. IRI can be seen as "URI or sequence of characters plus an algorithm"[3]. So even if authors don't care (and they shouldn't need to) and even if implementations do it right already, it needs to be specified. Even if there is just a link to [4].

Greetings,
Dirk

[1] http://www.w3.org/TR/2011/REC-CSS2-20110607/syndata.html#uri
[2] http://www.ietf.org/rfc/rfc3986
[3] http://www.w3.org/TR/SVG/linking.html#IRIReference
[4] http://www.ietf.org/rfc/rfc3987.txt

> 
> ~TJ
Received on Tuesday, 28 August 2012 22:39:57 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:59 GMT