Indeed, though if the authors are saying
> The number field sieve algorithm is still far from being fully understood, in particular for extension fields that are so important for pairing-based cryptography.
I won't say I understand it either. But it indeed seems to only be applicable in towers of extension fields which are only used for pairing-based cryptography.
