Re: Undefined behavior of missing feFunc[RGBA] in feComponentTransfer

Hi Robert,

This is an official response from the SVG Working Group.

Robert Longson wrote:
> In the errata text it says...
> 
> Specify that a missing feFunc[RGBA] acts as a no-op, and for 
> feFunc[RGBA] specified multiple times the last one is used (check that 
> this is backwards compatible).
> 
> Why choose the last one? Note that for multiple desc/title elements:
>

The SVG Working Group chose this behavior because it was consistent
amongst implementations of feFunc[RGBA] at the time of discussion.

> If user agents need to choose among multiple 'desc' 
> <http://www.w3.org/TR/SVG11/struct.html#DescElement> or 'title' 
> <http://www.w3.org/TR/SVG11/struct.html#TitleElement> elements for 
> processing (e.g., to decide which string to use for a tooltip), the user 
> agent shall choose the first one.
> 

The chances of someone relying on the behavior outlined in the errata is
quite small. Changing the behavior such that user agents choose the
first feFunc[RGBA] would make make the specification more consistent. It
would also require implementations to change their behavior. However, 
the change would be low risk as it will probably have little to no 
impact on existing content. It is considered to be a corner case for 
most implementations.

The SVG Working Group feels that this is an arbitrary decision either
way. If you feel that this is an important issue and that the errata
should be changed then the SVG Working Group is happy to discuss it in
future meetings and telephone conferences.

> Best regards
> 
> Robert

Thank you for your feedback.

Kind Regards,

Anthony Grasso

Received on Saturday, 3 November 2007 00:01:38 UTC