You appear to be citing photozone.de.
No, they are citing me. I have been critical about Sigma's bokeh before they published their review.
Perhaps you have, though not on this site.
In their images (e.g. of the reed grass), I don't see the "transition zone" issues, especially when compared with the corresponding images in the 35L test which are unarguably less pleasing. It doesn't strike me as a valid reason to disregard the Sigma.
Look at the image with the garbage bin in the foreground; the grass on the right of it. As I said, the Canon has a similar problem but not that bad, IMO.
We would need to see the same image shot with the Canon to make a comparison.
In the interests of balance, viewers of this video voted the Sigma's bokeh best in a blind comparison.
Battle of the Bokeh - Canon, Nikon Sigma 35mm f/1.4
This is far from the transition zone. Not challenging enough.
At least it has validity by being a comparison of the same view.
Here is a shot with problematic bokeh with the 35L, f/1.4 (my image): It is more or less a torture test. Look for the double lines there. If I wanted to convince you that the 35L had a wonderful bokeh, I would have posted a different image, like this one or this one (both f/1.6). But that would have been either cheating or lack of knowledge because the first one contains foreground blur only (which tends to look good with most well or over corrected lenses), and in the second one, the transition zone is small and does not dominate the image.
I understand the transition zone to be the transition from in-focus to out-of-focus, like the grass in the image of the rubbish bin. The busy background in this fairground image is in the distance.
So the bottom line is - it takes more than one or two samples to understand what a lens can and cannot do.
Indeed. But if one is criticising the bokeh quality of one lens against another, comparative images of the same subject are more reliable evidence than isolated samples.