++Æ can be freely used

I have formally withdrawn all the patent claims covering all the ++æ variants so far published either offically as CAESAR candidates (i.e. versions 1.1 and 1.2) or in this blog as  v2.0 candidates . Moreover, to the knowledge of the author, no one of them is covered by any other IPR restriction. Thus, anyone willing to further develop any of them, or to use, is free to do so (and welcome).

Cheers !

(to be continued?)


September 28th, 2014 Update:

Daniel Bleichenbacher raised a flaw in the proposed scheme for Associated Data: any triplet ... || x || y || y || ..., in the AD string can be replaced by ... || y || x || x || ... and the modification will not be detected. A silly mistake of me :(

I suppose it can be easily fixed somehow (in fact, I'm tempted to post some simple solutions) but I wont it do by my own. Working in a solo team and devoting only a fraction of my spare time makes me unable to consolidate a v2.0 for ++æ. In any case, I found some fun and beaty during this exercise with Fibonacci sequences applied to AE and I'll go on, at a slower pace and only for my eyes by the moment, with some funny concept in this line for a even much simpler AEAD scheme ...

October 21st 2014 Update

... just to write down that a generalized sound approach would be to handle the AD data in the same way that cryptogram blocks when processed by the authenticated decryption algorithm ... If i find some time I'll update some day the corresponding diagram specifications for the three ++a2 v2.0 candidates ...