W3C home > Mailing lists > Public > public-blockchain@w3.org > March 2017

Re: DNS Blockchain Use Case

From: Melvin Carvalho <melvincarvalho@gmail.com>
Date: Wed, 15 Mar 2017 14:47:40 +0100
Message-ID: <CAKaEYhJ15tPwDrN1revdgwq=muA_xiJX00=Lr1DjcW+TW-3kNQ@mail.gmail.com>
To: "Michael D. Palage" <michael@palage.com>
Cc: Blockchain CG <public-blockchain@w3.org>, Zooko Wilcox-OHearn <zooko@leastauthority.com>
On 15 March 2017 at 13:27, Michael D. Palage <michael@palage.com> wrote:

> Hello All,
> During the ICANN regional meeting in Copenhagen yesterday, the following
> presentation was made in a session of “Emerging Identifiers” which proposed
> a blockchain based domain name system
> See http://schd.ws/hosted_files/icann58copenhagen2017/44/
> NAMECOIN%20Rand%20ICANN58%20Emerging%20IT%20Session.pdf

Ah, the infamous Zooko's triangle strikes again! :) (cc'd

This one has been been doing the rounds for quite some time

Zooko's triangle, roughly speaking, says it's problematic to overload an
identifier to do three specific things.

I think that's missing the point slightly.

In good engineering it is better not to overload identifiers to do achieve
two specific goals, let alone three.  It may be convenient at times to
overload, but generally speaking, It's an anti-pattern, and comes at a
cost.  Particularly wrt interoperability.  This is one reason we have
fallen towards centralized web, rather than designing a nice mix of
central, decentral, federated and distributed technologies.

Good identifiers should do one thing well.  If you want to do two things
have a lookup mechanism, tying key value pairs to an object.

Namecoin is a lookup mechaism via a block chain, which was a clever idea.
But it's not a particularly secure block chain, it's merge mined and there
has known attack vectors.

There's better lookup systems, such as (at least on the web) HTTP which is
understood by almost all devices, and benefits from a large network and
install base.

Hopefully new tools such as semantic block chain will be able to illustrate
such techniques in solving problems in a more open ended, feature rich and
scalable way.

> Best regards,
> Michael
Received on Wednesday, 15 March 2017 13:48:15 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 15 March 2017 13:48:15 UTC