Hironobu SUZUKI wrote:
The license (if one can speak of a license at all, the page is not
clear that it is a license agreement) do not appear to be perpetual,
and it seem to restrict implementation to be unmodified Camellia.
I guess their intention is to protect their hardware version of
Camellia implementation because they build not only software version
but also hardware version of Camellia.
As far as I know, Japanese version of patent law is very limited on
Software. It is very different from US patent law situation. NTT is
a Japanese company and NTT speaks in Japanese context.
Hmmm... I think we would need a patent lawyer to opine on
what the ramifications are. Patent law is full of all sorts
of tricks and turns, it isn't like "ordinary" law where you
can use common sense.
I personally would be skeptical. Unless there is a clear
and open licence for anyone to use that algorithm, then
saying it is patented is probably the best description.
The problem is that if we start using this algorithm, and it
becomes popular, then NTT can change their minds and start
to go after people. Unless they have granted a clear
perpetual licence, this is something that is hard to protect
against. The guys at NTT may be really nice guys, but they
may not be there in 10 years time.
The problems that were had with IDEA seem to point us to
non-patented algorithms, simply because it is more certain.
We can get GPL, BSD License and others version of Camellia source code
that are provided by NTT that is Camellia's patenter. This Camellia
code is well tuned up and runs very fast. It is not sort of sample
code.
http://info.isl.ntt.co.jp/crypt/eng/camellia/source.html
If we can NOT use, copy, modify, and redistribute it, that is a sort
of license violation against GPL by the original licenser. I believe
NTT knows what GPL is.
The licence to the source code is orthogonal to the patent
situation. The source code is not the invention.
To some extent, they tried to address this issue in GPL v3,
I have no idea how successfully...
iang