MIME-Version: 1.0
In-Reply-To: <20140218214721.GA25356@savin>
References: <le05ca$qn5$1@ger.gmane.org> <5303B110.70603@bitpay.com>
	<20140218214721.GA25356@savin>
Date: Wed, 19 Feb 2014 01:41:36 +0200
Message-ID: <CAF7PVPq+g98J-Q8Mssp5ap9cfrhPmwh91E8qn4gSEiHPApgx5g@mail.gmail.com>
From: Bernd Jendrissek <bitcoin@bpj-code.co.za>
To: Peter Todd <pete@petertodd.org>
Content-Type: text/plain; charset=ISO-8859-1
Sender: "Bernd Jendrissek,,," <berndj@vps.pedantic.co.za>
Cc: bitcoin-development@lists.sourceforge.net
Subject: Re: [Bitcoin-development] BIP70 proposed changes
Precedence: list

[Ick, resending to list due to From: snafu]

On Tue, Feb 18, 2014 at 11:47 PM, Peter Todd <pete@petertodd.org> wrote:
> What specifically do you dislike about X.509? The technical standard or
> the infrastructure around it? (IE the centralized authorities)

I'm not the one who was complaining, but what I dislike is that a
certificate can have only one issuer. Cross-signing doesn't address my
dislike: it's different enough from being a certificate's single
issuer that it leaves too much power in the CAs' hands, IMHO.

It isn't so much the centralization per se that I object to, but the
way that the technical standard encourages concentration in the
infrastructure. See
http://lair.fifthhorseman.net/~dkg/tls-centralization/#Why_does_the_architecture_encourage_concentration%3F

I've been (slowly) working on a patch to allow pki_data to contain
more than just the single certificate chain that the
single-issuer-only format insists on, but I'm making as many steps
back as forward, being unsure of the right way to do it. Implementing
an OpenPGP-based pki_type would probably be better, but hacking x509+*
seems like a lower-hanging fruit.