Currently, I have 4 50s, Olympus Zuiko 50 F/1.4, Contax Zeiss T* Planar 50 F/1.4, Canon EF 50 F/1.8 Mk II, Canon EF 50 F/1.2L, and I tested so many 50s by now I start to lose count . So yes, to me it is an important lens, both on APS-C and on FF, as it is a very nice short portrait tele on APS-C, and because it is a great all purpose lens on a FF.
I also think that for low light photography, it is nice to have a 24/28 - 50 - 85/100 setup in fast primes. I started off with 28 F/1.8, 50 F/1.8 Mk I (after testing 3 50Ls, 4 50 F/1.4s and 5 50 F/1.8s) and 85 F/1.8. The 28 was for (small) group shots on 40D, and larger groups on 5D, 50 for head and shoulder shots on 40D, small groups on 5D, and 85 for tight head shots on 40D and head and shoulder shots on 5D. This means getting (quite) close, for which the 50Ls I tried didn't work too well at the time.
I replaced the 85 F/1.8 with an 85L, and the increase in IQ at large apertures meant the 85 F/1.8 just didn't get camera time anymore, so I sold that. The 28 was replaced with the 24L, although I still miss the 28 because of its IQ and compactness, and eventually I managed to get a 50L which works well.
As I mentioned, I use these as low light lenses, and I often shoot these at F/1.2 or F/1.4 to F/2, at 1600 and 3200 iso. However, I also use them for other purposes, f.e., landscapes, at any aperture, and the 50L as a general walkaround on the 5D, the 24L as a general walkaround lens on the 40D.
As mentioned, I tested and tried quite a few different 50s, and have come to the conclusion that the 50L is the best of all, provided you get a good one, for colour, sharpness, contrast and bokeh, at all apertures, where F/1.2 is a little soft, but very usable, F/1.4 - F/1.6 is great, and from there onwards it is excellent.
I will likely try other 50s when I come across them, but my general conclusion so far is that F/1.7s/F/1.8s/F/2s generally are good from F/2, and are sharpest at around F/5.6-F/8, F/1.4s are good to excellent from F/2.8, and sharpest from F/4 to F/5.6, and that the 50L is in a league of its own. From F/2.8 some F/1.4s are similar or come close (Sigma/Canon when they focus well, and the Contax Zeiss, with the Sigma being very sharp from F/2) but at large apertures it is IMO the lens to beat.
I think it is about time manufacturers redesign their 50s, because almost all are old school Gaussian designs, which is why they seem to perform very similarly, with the exception of the Sigma and the Canon 50L. They all seem to be much sharper in the centre until you stop down a few stops, when they become more or less equally sharp in the corners. The F/1.7s - F/2s are better up to F/2.8, probably because they use smaller glass, the F/1.4s tend to beat the F/1.7s to F/2s from F/2.8, probably because the slightly better designs exclude the worst parts of the glass at F/2.8. An F/2.5 or F/2.8 50 mm tends to be excellent from maximum aperture, BTW (some of the macro lenses), provided it is a Gaussian design, and not a triplet or Tessar-type design (Jena Domiplan 50 F/2.8 comes to mind in the latter category).
Regarding the Sigma 50 F/1.4 and the 50L: both have been designed with bokeh in mind, and as a result have residual focus shift when stopping the lens down from maximum aperture to around F/2.8 or F/4 when focusing at close distances. At its worst this may well be about 4 cm, which means that focusing on an eye may yield sharp temples rather than sharp eyes, which is a pain, obviously. Not all copies show this in such a bad way, as this shift may occur within DoF and hence not be all that noticeable, but it looks like it is part and parcel of the deal for great bokeh in these lenses.
The two Sigma 50 F/1.4s I tested both had their worst shift at F/2, about 1 cm at a focusing distance of around 50 cm, which is outside DoF at that distance and aperture, the 50Ls I tested varied from 1 cm to 4 cm at a similar distance, except for the last two of a total of 5, which seemed to have only some backfocus (2 cm and 4 cm respectively). The 2 cm backfocusing copy was calibrated twice by Canon, and the second try made it work like it should. The Sigma I had never calibrated, these two also had problems focusing in low light at longer distances. I didn't check whether this was caused by focus shift or not.
Anyway, so far some of my thoughts.
Kind regards, Wim