Well, of course certain people in the UK aren't too happy with the harmonised rule removing the old adverse Xwind approach.
Personally, I have no great dispute with it; I'm not aware that non-CAA aircraft suffer significantly more runway excursions, so regardless of the theoretical merits of the CAA method, the practical effect doesn't seem to have been to increase safety.
Now one might argue that using the historical data may be less valid now, when the test techniques are more sophisicated, and we're probably shaving more margin off in other places as we "learn more". There's a world of difference between DGPS and a guy dropping a bag of chalk to measure deviation, after all.
But it's all we have to go on.
There's other compromises in the Vmcg tests, too, other than Xwind. Wet runway would be another issue; nosewheel steering off removes the steering, but the mainwheels are still working on a dry runway.
I think the way we do it now is probably as defensible as any other; you could argue, for instance, that in a Xwind landing the pilot may be more active/aggressive in heading control, and perhaps faster to react to the initial heading swing; that makes a huge difference, of course, to the deviation.
In fact, tests in a (mild) Xwind might even be artificially improved; if it's a bit gusty too, there's a risk that the pilot is applying a input in the "right" direction when the engine is failed, and your deviation might be optimistically low. At least low wind conditions are going to be consistent (as much as Vmcg ever can be!)
-----
Regarding Vmca, there are issues with those tests too; it's common to test at lightest practical weight, as that is the classical limiting case. But I believe I saw a paper from Airbus some years ago that suggested that alpha effects on lat-dir aerodynamics could be important (among other concerns), which might make a heavier weight more critical for the same CAS.
-----
And in the case of both tests, the way the industry approaches the thrust to use for test is a bit odd too; the guidance material isn't exactly logical when compared to the regs. It makes ASSUMPTIONS about the design when stating the test method, which can lead to contradictions between the intent of the regs and the method in the guidance. Much debates about whether it's valid to follow the letter of the guidance, even if it doesn't (strictly) meet the reg, since the guidance is stated to be a means to comply with the reg......
Regarding the CRJ VMC's. I don't think any Canadair/Bombardier products use alt/temp scheduling; it's only an option, after all. Why don't we? Well, on the older aircraft Vmca is declared as not limiting, so there's no real point in doing the extra extrapolation and/or testing. I think the merits for doing so were debated on the 700, but it must have been of limited perf value, because as you say it's a declared single value.