Google, please hire this person to rewrite the entire Keyczar documentation wiki. Keyczar might be great from a crypto point of view, but the lack of good documentation and binaries is probably a major reason why people ignore it and implement their own "dangerous" crypto using BouncyCastle.
While I'm making unreasonable and arbitrary demands, password-based encryption support in Keyczar would be great too, thanks.
BouncyCastle is a low-level library, like OpenSSL. It's very hazardous to use in practice. It also gets less cryptographic scrutiny than OpenSSL does.
The point of this article is that you should use high-level libraries that supply high-level constructions you can use directly in your code, rather than primitives that you'll have to knit together into high-level constructions yourself.
Keyczar is a wrapper around crypto primitive libraries. The .net Version actually is based on Bouncy Castle even. However the big downside is that Keyczar is not maintained well. Pull requests sit for years, not good for a security related framework.
2015 and Google still can not be bothered to put their stuff in Maven Central. It is as if they were actively trying to stop people from using their stuff.
While I'm making unreasonable and arbitrary demands, password-based encryption support in Keyczar would be great too, thanks.