Re: [bitcoin-dev] Scaling Bitcoin conference micro-report

["Eric Lombrozo" <elombrozo@gmail.com>]

------ Original Message ------
From: "Milly Bitcoin via bitcoin-dev" 
<bitcoin-dev@lists.linuxfoundation.org>
To: bitcoin-dev@lists.linuxfoundation.org
Sent: 9/20/2015 3:51:36 PM
Subject: Re: [bitcoin-dev] Scaling Bitcoin conference micro-report

>>Some of us have also been actively working towards developing
>>a more modular, layered architecture and better implementations that
>>will afford greater decentralization in software development with less
>>need for critical code reviews, less pushback from downstream 
>>developers
>>who must continuously rebase, a better process for building consensus 
>>in
>>the community, and simpler app migration.
>
>It sounds more efficient but it is not clear to me that it would change 
>the level of centralization of how the final decisions are made.
>
>One threat to Bintcoin involves incentive for companies to hire 
>developers.  The only reason is to change (or not change) Bitcoin Core 
>so it is beneficial to their interests.  I am not sure anything can be 
>done about that risk but it needs to be understood and considered and 
>not just ignored.
Core development process and decentralized dev/community consensus 
building (in particular for consensus-critical changes) is at the top of 
my priorities as issues right now...and one that I'd love to discuss 
more in depth...but it probably deserves its own thread. The political 
angle seems very difficult right now while the systems architecture 
stuff seems a bit more tractable...and it seems that without 
architectural changes it will be extremely hard to decentralize 
development and easily bring large numbers of new developers in.

>
>>We need to increase the basic infrastructure nodes by a factor much
>>larger than 2 or 3...more like 100 or 1000...and it's entirely doable
>>with properly aligned incentives.
>
>I assume that would mean fees that hike transaction fees and make 
>Bitcoin more expensive?
>
Not necessarily. Right now we already pay around 3,600 bitcoins a day in 
inflationary subsidies, very little of which goes to the majority of 
critical infrastructure nodes and their operators. This is a problem 
with the current protocol design, one we'll hopefully be able to fix.

Having more core infrastructure nodes doesn't need to raise costs per 
transaction - but it will most likely require abandoning the current 
approach of having three basic node classes: miners (which tend towards 
centralized pools), full nodes (which must validate each of everyone's 
transaction and in return get paid nothing), and thin clients (which 
essentially amount to parasitic nodes that do not contribute any 
resources back to the network and must be subsidized).