Re: [http-state] http-state charter

[Adam Barth <[e-mail]>]

On Tue, Aug 4, 2009 at 12:14 AM, Daniel Stenberg<[e-mail]> wrote:
> On Mon, 3 Aug 2009, Adam Barth wrote:
>> The "Informational" status does not appear appropriate for this document
>> because the Internet community seems to have reached consensus about the
>> behavior of the Cookie and Set-Cookie header as evidenced by billions of
>> such headers traveling over the Internet between hundreds of implementations
>> every day.
>
> As on author of one of those hundreds, I for sure would've appreciated the
> existance of such a document when I started my work.
>
> If clarifying the existing practise isn't what you want to see, what exactly
> do you want?

That is what I want to do.  :)

More specifically, I think we should document behavior that is broadly
interoperable.  For behaviors that differ between major
implementations, we should use our judgement and spec the most sane
thing.

> Both RFC2109 and 2965 failed because they tried to engineer a new cookie
> concept that no author was interested in. I don't see how craftning yet
> another new thing has any better chance this time, so the "only" remaining
> work is instead documenting how things actually work today.
>
> Or what am I missing?

I think we should obsolete those cookie specs and replace them with a
standards track cookie spec that is actually useful.

Adam