August 18, 2014 Status and Thoughts.I have about half of the Texai cognitive architecture migrated to GitHub. The knowledge base repositories are larger than GitHub guidelines, so I need to find another more appropriate cloud data server, e.g. Dropbox for those. I configured three separate computers in my lab with the same development environment to ensure redundancy and to document the build recipe. The configuration for developers is . . .
- Ubuntu 14.04 LTS on bare metal
- Oracle Java 8 with JCE highest level security policy
- Netbeans 8.0 Java IDE
- Maven 3
- Docker
The Docker platform is currently used to host a Jenkins continuous integration server that I uploaded to the Docker container registry.
My GitHub page is
https://github.com/StephenLReed?tab=repositories.
My Docker repository page is
https://registry.hub.docker.com/u/stephenreed/jenkins-java8-maven-git/.
I would like to have bitcoind in an agent-controlled, non-verifying proxy by early October, 2014, for hallway demonstration at the Hashers United conference in Las Vegas. Consequently, I will upload broken aspects of Texai that have no relationship to Cooperative Proof-of-Stake, as the agents I need to write are dependent upon working code embedded in those otherwise broken dialog modules.
------------
Here are my latest thoughts about how CPOS will work. I am considering hard-forking the popular Satoshi blockchains: Bitcoin, Litecoin, Dogecoin and Namecoin. They would operate on networks separate from the Satoshi originals and be branded respectively CPOS-B, CPOS-L, CPOS-D and CPOS-N. Users would be required to configure their respective client wallets, payment processors, mixing services, etc., to respective CPOS seed node addresses - that's all. There are no changes required to the client protocols I believe.
The forks could come much earlier than January 2016, depending upon the ambition of the network and the state of system testing.
I would use a portion of block rewards to pay geographically dispersed operators to run full nodes for these plus a Tor relay. Each operator would be required to provision a high-availability enterprise server in a datacenter having redundant high bandwidth Internet connections. Specifications for the server will include protection against DDoS attacks.
As operators will be relatively well compensated, I expect to add new servers only with increased market capitalization and transaction traffic. VisaNet runs the majority of the world's credit card transactions with only two datacenters, so less than ten geographically distributed servers with likewise separate ownership may be adequate to start. Supposing that CPOS grows to the same market capitalization that Bitcoin has currently, then 10,000 distributed paid servers could be justified.
I would choose either CoreOS or Ubuntu LTS for the base operating system, and use Docker containers for the cryptocoin bitcoind software, where each coin gets its own container. Likewise the tor relay gets its own container. Each server would run the latest identical software and as the network grows in size, there are existing softwares to manage the worldwide deployment of Docker containers. Unpaid volunteers could download and run the CPOS agent software to record blockchains and verify peer agent behavior, but would not be permitted to operate agents that mint new blocks or that provide support services to other agents.
Governance of the CPOS network will be performed by the Texai cognitive architecture, in which simple transparent algorithms enforce policies decided at periodic meetings of the node operators. Each operator will be required to purchase a commercial membership in their local Bitcoin Foundation. A portion of the block rewards will fund developers tasked via the Texai cognitive architecture. Each paid person performing a task for the system will have at least two other non-affiliated paid persons check their work. A portion of the block rewards will fund development of the Texai architecture, Bitcoin Core, and relevant altcoin features. Supposing that CPOS coins grow to the current market capitalization of bitcoin, then 300+ developers could be supported. Paid human participants in the CPOS system, e.g. operators and human agents, could be anonymous but would be required to possess X.509 certificates for digital signatures and network authentication. Likewise they must provide verifiable contact information. All communications in the network are digitally signed, TLS/SSL encrypted, and logged. We want the network to be trustless in the sense that Satoshi's Bitcoin is trustless, but paid participants must be held responsible for their behavior via cyrptographic credentials and tamper-evident logs.
I expect the block rewards of the CPOS coins to greatly exceed the transaction revenue with respect to paying for the system. Consequently, I would set transaction fees to the lowest sensible values which I believe are 100x less than what bitcoin charges. Rather than $.05 fee at today's bitcoin price, I would set CPOS transaction fees at the equivalent of $.0005 which encourages the use of microtransactions.
As an example of "eating your own dog food", the Texai software agents will pay each other for services rendered using bitcoin microtransactions. This will encourage software developers to write skills for the system, as they will receive the net earnings of software skill they provide in accordance with a scheme to maximize innovation.
In accordance with the precedent set by Satoshi's Bitcoin, the CPOS network will not be a legal entity in any jurisdiction. The network, despite Tor anonymity capabilities, will not place servers in jurisdictions where their operation would be against the local law. The network will pass through all earnings, e.g. block rewards, transactions fees, and software agent income, to human participants who will pay taxes according to their local jurisdictions.
