Instead of hard-coding the desktops, I guess we should link to pages like this one:
https://docs.fedoraproject.org/en-US/releases/f42/blocking/

Nah, there's a distinction there. Blocking desktops and blocking images are not the same, though obviously related. Note for instance that we don't block on the Workstation or KDE lives on aarch64, even though Workstation and KDE are blocking desktops.

The criteria are the canonical definition of what desktops are blocking. The blocking images page is the canonical definition of what images are blocking.

We should just update the page, if we do now block on KDE on aarch64 (when/where was that decision made? A reference would be good.)

We should just update the page, if we do now block on KDE on aarch64 (when/where was that decision made? A reference would be good.)

Here:
https://fedoraproject.org/wiki/Changes/Fedora_KDE_AArch64_ReleaseBlocker

Blocking desktops and blocking images are not the same, though obviously related.

So, why don't we specify just GNOME and KDE as release-blocking desktops in criteria, and make the architectures be visible from the blocking images page, instead of hard-coding them?
(I do admit that the difference is minor, and I don't really care either way we do it, I'm just curious whether I still misunderstand something).

We have release-blocking desktops and arches specified in the criteria, and then we have specific images listed on the images page. An image can be for a blocking desktop and a blocking arch but still not be release blocking in itself - that's the point I was trying to make above. This is the case with the Workstation aarch64 live image, for instance.

I've adjusted the wiki:
https://fedoraproject.org/w/index.php?title=Template%3ARelease_criteria_preamble&diff=734730&oldid=655291