[Cryptography] Standards Trolls: Re: Bitcoin is a disaster.

[Greg]

> This is actually, in my view, a definition of a *distributed* system and is not what I think people normally mean by "decentralised" (but it's hard to tell because when I push on that particular point people generally stop talking to me, or claim that it is "obvious" and I should stop wasting their time).

As far as my research has shown me, the term “distributed” comes from “distributed systems” and “distributed computing”, which are terms you are free to look up even in places like Wikipedia, and neither of these terms discuss decentralization, which is a term whose origins come from as far back as the French revolution (at least according to Wikipedia), referring to the decentralization of power.

The same holds true today, that in most conversations, discussions, lectures, research papers, etc., the terms “distributed computing” have almost always referenced systems where computers that are all owned and operated by a single entity (read: centralized), are distributing their compute resources across multiple machines.

> This surely cannot hold for a truly decentralised system since such a system can surely remain partitioned forever.

Sorry, not quite clear to me what you’re saying here.

Cheers,
Greg


> On Jan 7, 2021, at 4:45 AM, Ben Laurie <ben at links.org> wrote:
> 
> 
> 
> On Tue, 5 Jan 2021 at 02:08, Greg <greg at kinostudios.com <mailto:greg at kinostudios.com>> wrote:
> See the DCS Theorem [1].
> 
> This uses an extremely weak version of "decentralised" (but it least has a definition, which is far from the norm). That is:
> 
> "Decentralized means the system has no single point of failure or control (SPoF). Another way to state this is: if any single element is removed from {𝑆}, the system continues to perform its intended behavior, and no single component in {𝑆} has the power to redefine 𝑓𝑆 on its own"
> 
> This is actually, in my view, a definition of a *distributed* system and is not what I think people normally mean by "decentralised" (but it's hard to tell because when I push on that particular point people generally stop talking to me, or claim that it is "obvious" and I should stop wasting their time). This is illustrated by a point made in the proof:
> 
>  "for a system to be considered decentralized, it must be uncompromised, and that in turn means it successfully processes all authorized messages from new users within some interval 𝑆𝑡 ."
> 
> This surely cannot hold for a truly decentralised system since such a system can surely remain partitioned forever.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20210107/918dedce/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: Message signed with OpenPGP
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20210107/918dedce/attachment.sig>