Imho, either we can get rid of all O** encodings, including maths, or
it's only a not very usefull way to overload ctt with cries.

And the t1 (type 1) issue is quite relevant, here.

I have been using default T1 formats for years, and switched back to
OT1 by default because users didn't understand why their source files
were not properly treated elsewhere...