One important question: how do you manage GFDL on spoken text? To the satisfaction of, e.g., querulous Commons admins who deal with licensing stupidities all the time? (Geni, I'm looking at you ;-) )
Requiring a reading of the license on the end of all audios is onerous. Our many spoken articles on English Wikipedia are (presumably) not a violation as long as they're on Wikipedia, with the license text a link away - but aren't really unencumbered for use elsewhere.
Is the GFDL fundamentally discriminatory against the blind?
Kat Walsh has asked licensing@fsf.org, but they tend to act like a Magic 8 Ball that says "read the license text and consult your lawyer."
I asked about audio versions of GFDL text on the FSFE discussion list. One useful suggestion (from M. J. Ray) was:
Not in England if done to allow access by visually impaired people in certain circumstances (Copyright ... Act 1988 sections 31A-31F). There's probably other special cases too.
Is there such a provision in US law? I presume there is one in other legal systems too.
This in itself IMO is a strong case for porting to CC-by-sa.
- d.