On Wed, May 31, 2017 at 2:59 PM, Dave Cridland <d...@cridland.net> wrote: > But you did copy, amongst possibly other things, a constant-time > conditional-swap routine, along with the comments, from the > "additional" directory of a GPL library, so I can only assume that > copyright law works an entirely different way in this parallel > universe. Is it like patents in my universe, where a defence is that > the invention is obvious to one skilled in the art? We don't have that > defence for copyright in my universe.
Yes, I copied their cmov implementation. This is a basic operation that be found in other libraries and will always look roughly the same; there are only so many ways to write a+b. In fact, an implementation for a different data type was already in the Go library (https://github.com/golang/crypto/blob/master/ed25519/internal/edwards25519/edwards25519.go#L61), so I could have just as easily copied that and it would have come out the same, or copied the ref10 implementation (which that function was adapted from) for fieldelements which is the same thing (but for an array over a struct), which is probably what OWS did. Or just gone on wikipedia, I'm sure they have pseudocode for a conditional register swap that I could have stuck in a loop. OWS did not invent or even really implement this; it's just what the operation looks like. —Sam _______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: standards-unsubscr...@xmpp.org _______________________________________________
