I did not say "intended" to undermine it. It just does on account of the burden of managing trees of attributions.
In academia we attribute anyway even for old PD works.
The 98% is not a valid reason to drop CC-SA. So, yes, we should lobby for Attribution to be an option and reinstate CC-SA.
Right now CC-SA 3.0 is not an option (retired at 1.0), so the percentage of people using it is not likely to grow (as I believe it would).
Insisting on "BY" reinforces the notion of "IP" which has been contended in many places (e.g. here).
Keep up the good work towards a libre curriculum :-).
There is still a 1.0 version of the CC-SA license that one could theoretically use, but since that would cordon off all of that material in yet another silo where none of it could be remixed with BY-SA or GFDL material, I think that would be a big mistake. Better to simply release it into the public domain or otherwise under a "no rights reserved" license.
Thanks Steve for the comment. Agreed, it does not make sense to use the CC-SA 1.0 license - for the reasons you mention and also because it is out of date and lacks some of the requirements included in the 3.0 licenses (e.g. relating to DRM) - hence the "share similarly" approach suggested here.
Apparently, in Europe it is not legally allowable to release your work directly into the public domain. The only way for one's work to be in the PD is for its copyright to expire.
Do you know of any "no rights reserved" licenses? As far as I can see there is still a need for a license that would be more or less equivalent to PD.
This Libre License is also trying to make a stronger statement about consistency right down to the requirements for free data formats and free software for manipulating and sharing knowledge resources. It is more about a shared set of values around inclusivity: see Declaration on libre knowledge.