Hacker News Reader

Comment by _stephan | Hacker News Reader

Congratulations!

Who knows, maybe a part of the "exciting plans for 2015" is to release PolarSSL under a liberal open source license...

The only advantage of doing so would be to allow small closed-source, proprietary shops, who can't currently afford the €99/month subscription, to use the library. I struggle to sympathise in this category, it's not like this group is locked out of SSL/TLS entirely.

Ah, most noticably, it'd be an interesting alternative for BSD/MIT (and Apache?) proyects including the *BSD family of OSs.

It's GPL 2 - https://polarssl.org/how-to-get#open-source

I'm not saying that's the nadir of licenses - but it's pretty good, no?

GPLv2 crypto will never dethrone OpenSSL

GnuTLS is LGPL and can be used alongside commercial applications. There's also Botan (BSD, C++), Bouncy Castle (MIT, Java).

Who should care if GPL is bad for a SSL library? First of all it needs to be correct and PolarSSL is the only one in this field. It is trusted, and it is fast enough.

I would never trust GnuTLS, NSS or OpenSSL over PolarSSL.

> Who should care if GPL is bad for a SSL library?

An entity which actually ships software?

> I'm not saying [GPLv2 is] the nadir of licenses - but it's pretty good, no?

It's certainly "pretty good" compared to having no source code available at all. But when people talk about "liberal" licenses (aka permissive licensing), the GPL is not what usually comes to mind.

This is ARM we're talking about. They're not known for being open-source friendly.

ARM fund CodeSourcery to get ARM compiler support into the gcc toolchain. Plenty of people there like open source.

The worry, internally, is a holdover from the days of the SCO-Linux disputes - that if an employee works on an open-source compiler, sees a great idea, then remembers it and implements it in the closed-source compiler later on (even subconsciously) that it could mean a bunch of legal hassle.

Reply on news.ycombinator.com