Re: [TLS] drop obsolete SSL 2 backwards compatibility from TLS 1.3 draft

[Brian Smith <[e-mail]>]

Dave Garrett <[e-mail]> wrote:
> On Saturday, December 27, 2014 07:39:07 pm Brian Smith wrote:
> I'm hesitant to even mention it, but obviously my next most preferred route is
> obviously to do that and make acceptance of an SSL2 hello for older TLS versions
> a SHOULD NOT rather than a MUST NOT.

That seems good to me.

> That, however, is basically just RFC 6176 with a wag of a finger thrown in.

That's good, because it should be easy for everybody to accept.

> There have been bugs caused by supporting this and it should be mandated to no
> longer be usable for all implementations. Anyone actually relying on it and not
> capable of flipping the client-side switches needed to stop it is not likely to
> be a candidate for early adoption of TLS 1.3, anyway.

Servers generally cannot flip switches on the client.

I want TLS 1.3 to be deployed widely and quickly after it's been
standardized, and I want people to be able to put it into production
use during the standardization effort so we get the same kind of
interop testing that HTTP/2 got. This requires that the TLS 1.3 spec
should try to avoid breaking how earlier versions of TLS work, unless
there is a very strong reason to do so.

Cheers,
Brian