Some thinking while I write...
I haven't done any research at all on this, just musing.
My Thoughts were to grow the purple CMS carrots next to my fertile short fat purple (SFP) carrots that I'm trying to stabilise.
Any seed from the purple CMS carrots will have been pollinated by the fertile SFP carrots, and hopefully inherit the SFP characteristics, while still retaining the CMS.
Next generation, grow the offspring from the two lines in parallel again, searching for SFP carrots in both lines.
I've just realised the flaw in this - the lines will eventually share much of the genetic material. So the chances of getting hybrid vigour will be reduced with each generation. At least in the nuclear genetics... I could split my SFP fertile carrot lines now, grow one line in isolation while doing the parallel growouts at some physical distance, and hope for genetic distance between the populations as well. Hmmm.
By chance, the CMS purples are growing next to a totally unrelated grex of short fat orange carrots - this would give a distinct genetic pool to draw from - but would also add a number of years to my project as I tried ro start selecting for SFP again...if I kept the lines distinct from my already stabilising line, I might get hybrid vigor out of a future cross between the new CMS SFP carrot , and already existing one.
So, then offer two distinct packets shipped together, one CMS SFP, along with my original ertile SFP carrot, so growers could make and maintain their own F1 SFP carrots - that would be CMS.
...There might even by more than one person in the world interested in doing it

G