Jan, I'm not a professional lens tester, but some properly applied imaging principles take one a long way and would suggest those photographs you offer above are not ideal for testing a lens. The criteria for the subject matter should include straight lines, sharp edges,
The photograph (not photographs, it's just one) has the straight lines and sharp edges necessary to show the CA.
fine surface detail, and a wide range of tonal variation all sufficient to evaluate distortion, egde rendition, detail rendition, and handling of contrast from corner to corner and border to border.
I quite agree. My test was more of an informal one; because the CA was an issue that bothered me in other
photos, that's what I wanted to show. The purpose of posting it here was to show that CA can
appear within the sweet spot range of f/8 to f/11 at 28mm, even with a DX-sized sensor.
That is how I selected the subject matter for my comparison test published on this website. Also blowing an image to 200% is not useful - at that magnification it becomes difficult to distinguish between lens quality issues and pixelation. Your testing material should be good enough to reveal what you need to know on your monitor at discrete intervals of 25%, 50% and 75% magnification, at which values aliasing of the monitor image and pixelation do not obscure inherent image detail. An A3 print at no less than 240 PPI native resolution should do the same in print.
As I mentioned, it is noticeable even in an A4 print; there is no need to print to A3 to see more of the same. In an A5 print, the CA appears as slightly blurry edges, but you probably have to look closely to notice. Of course, that's what people tend to do with such smaller prints.
While I agree with your criticism of the methods used, there is very little I can do about it a year later; as I stated, I no longer own this lens, and I can't do methodical testing with it.
If you're really interested, I'll fetch the original and post various magnifications (100%, 75%, 50%, whatever is your heart's desire ...).