W3C home > Mailing lists > Public > www-svg@w3.org > April 2010

Re: Margins for filters?

From: Jasper van de Gronde <th.v.d.gronde@hccnet.nl>
Date: Sat, 17 Apr 2010 17:20:00 +0200
Message-ID: <4BC9D1A0.6070803@hccnet.nl>
To: Erik Dahlstrom <ed@opera.com>
CC: www-svg@w3.org
Erik Dahlstrom wrote:
> On Sat, 17 Apr 2010 10:45:28 +0200, Jasper van de Gronde 
> <th.v.d.gronde@hccnet.nl> wrote:
>> ...
>> So, is there some way to specify a "margin" for the filter effects 
>> region?
> 
> There is in the SVG Filter 1.2 working draft (attributes 'mx', 'my', 
> 'mw', 'mh'), for details see the latest editors draft: 
> http://dev.w3.org/SVG/modules/filters/publish/SVGFilter.html#FilterPrimitiveSubRegion 
> 
> 
> Feedback on that would be most welcome.

It looks okay, but I was wondering whether it indeed makes sense to use 
the margin in addition to the normal filter region (apart from 
practicalities). As far as I can tell you would be able to "emulate" any 
old filter effects region using x="0" y="0" width="1" height="1" and 
some margin. So, you could consider letting the margin settings override 
the normal x,y,width,height properties. This way you could create a 
document that uses one filter region (the default for example, or one 
that is inconveniently large, or too small, or...) when rendered as SVG 
1.1, and another (that perfectly fits) when rendered as SVG 1.2.

Furthermore, have you considered an "auto" setting? For a lot of 
primitives it is trivial to determine the appropriate margin (they don't 
need one), and for things like blur, offset and the morphology filters 
it's not a big problem either. There are some (like convolution) that 
could be a little bit trickier (when using the bias property for 
example), but I think most can simply do without a margin by default.
Received on Saturday, 17 April 2010 15:20:37 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:44 GMT