Re: ISSUE: Datatype for feTurbulance 'seed' attribute

Thomas,
Thanks for reporting this problem.

I checked with Adobe's implementers and our conclusion is that you are 
correct. I agree that an SVG 1.1 erratum makes sense. I raised a formal 
issue within the SVG WG on this so that it won't fall through the cracks 
and the WG can decide if they agree with our conclusions.

Jon


At 05:14 AM 9/21/2005, Thomas DeWeese wrote:

>Hi all,
>
>    It recently came to my attention that the SVG 1.1 spec
>states that the 'seed' attribute on feTurbulance is a
>'<number>' however the implementation provided in the
>specification requires the seed to be an '<integer>'.
>
>         http://www.w3.org/TR/SVG11/filters.html#feTurbulence
>
>    I suspect that the use of '<number>' is an editorial
>mistake and it was intended to be an '<integer>', if so I would
>appreciate if the WG would produce an errata to that effect.
>
>    If however the WG decides that this was not an editorial
>mistake then they should at least clarify how the floating
>point seed should be converted to an 'integer' for use in
>the code (rounding, truncate to zero, truncate to negative
>infinity, etc) - so that implementations can provide the high
>degree of interoperability required for this filter effect.
>
>    Thanks for taking the time to review this issue.

Received on Wednesday, 21 September 2005 18:16:40 UTC