Re: Some filter effects specification review comments

On 6/17/11 4:34 PM, "Erik Dahlstrom" <ed@opera.com> wrote:

>On Fri, 17 Jun 2011 15:02:32 +0200, Vincent Hardy <vhardy@adobe.com>
>wrote:
>
>...
>> Hit testing
>> 
>>https://dvcs.w3.org/hg/FXTF/raw-file/tip/filters/publish/Filters.html#fil
>>te
>> rs-in-css
>>
>> The filter is defined to have no effect on hit testing when clicking
>>into
>> the final rendered image. This is a concern: highly distorting filters
>> will basically create situations where the hit testing is completely at
>> odds with what the viewer sees. For example, in an extreme case, a
>>filter
>> can blow away the image entirely, but it will still respond to hit
>> testing. Has the group discussed possible solutions to this problem?
>
>A concern that's been mentioned before is that it can be expensive to
>evaluate filters for hit-testing. However, just taking the region that
>was  
>affected by a given filter could work performance-wise. Unfortunately
>that  
>region can be visually empty too depending on the input and what
>filterops  
>were used (e.g make something almost fully transparent).

Yes, and it should probably be looked at with pointer-events in mind too
(visiblePainted, visibleFill etc...).

I proposed this as a topic for the F2F on the agenda page:

http://www.w3.org/Graphics/SVG/WG/wiki/F2F/Seattle_2011/agenda_proposals#FX
TF_topics

Cheers,
Vincent

Received on Monday, 20 June 2011 09:08:27 UTC