Presumably such an extended time process works OK for b&w as Ilford mentions the use of it for several of their developers but the differences are so marginal in the case of b&w compared to colour negatives that 100% of viewers or as near as damn it would not see any difference?
Yes, and with color films, the differences would also be unnoticeable to many, because indeed...:
Finally the problem usually cited as the "intractable one", namely colour crossover, can be corrected via a hybrid process but cannot via an optical one? So "hybriders" can risk this extended time process and "cure" the crossover if it happens but not darkroom optical printers?
This is certainly the case. And to an extent, crossover can be dealt with in optical printing in a few ways that can make it less of a problem:
* Pre/post flashing to alter highlight color balance
* Supplemental masking
* Most importantly: embracing the crossover as a desirable feat of analog unpredictability.
As to the technical aspects of reuse of chemicals: I personally don't see ways to make this perform as consistently as properly replenished (high volume context) or one-shot use of chemistry (low volume context). The reason for this is that when reusing the developer, one can account for the reduced activity by extending development time, but this does not entirely account for the differences in color balance that result from halide buildup. Further problems arise from the small amount of sulfite being oxidized, and this in particular has a pronounced effect on dye formation - the effect will be counter to the reduced activity due to halide accumulation and developer oxidation, but again, this cannot be assumed to be equal across the color channels.
The effects of this are summed up quite well in Kodak's Z131 pub, especially page 5-32 which deals with developer replenishment issues. Note that not only overall density buildup is affected in an underreplenished developer (this, we can compensate for with extended development), but the color balance also is affected, with a pronounced difference between low and high densities (i.e. crossover).
In my mind, the main reason why people 'get away' with reusing non-replenished developer is because they simply don't notice the problems, and one might argue that they're not really problems for them. It all boils down to how strict one sets their criteria, and that's ultimately an entirely personal choice.