Recursive coordination portfolio


Portfolios are really effective at managing bets and risk. You own a little bit of everything to return as much as you can. What if we could manage our coordinations with others as a portfolio. People could decide to change their behaviour to benefit someone else. Such as writing documentation how some of their code works. Or adding an API


Everyone announces what they ARE doing.

Everyone comments on everybody else's what they are doing and says what they could do to benefit me. How to do that thing you're doing in my interest.

If the person doing the thing accepts that work, they add it to their portfolio.


(suppress notifications) (Optional) Please, log in.

Wait for a bit. I think, I'll make projects results notifications go along with comment notifications, so that we could engage with this process, while sharing work results where they belong -- under particular projects.

I've been looking at load balancing loops but my second idea is to load balance virtual processes A virtual process is a series of loops some nested. Software can be created as a series of processes. I want to learn how to schedule processes and consequently loops. I need to call .tick on each loop, in an order to advance its execution. I want the processes to be of different priorities and proportionate to eachother I might execute 2 loops of one process for every one of another. I also want to schedule .ticks for processes that depend on eachother

I found out how to implement intelligent DNS.

You can use PowerDNS with the pipe backend and to dynamically resolve a DNS query using the requests IP address. This could find servers that are near you.

Then you can health check Haproxies to find a load balancer that is healthy. That is intelligent DNS.

I am now thinking of scalability, specifically scalability using open source software. I was on a project that provided a reliable application across 3 availability zones with 3 F5 load balancers with DNS load balancing. The F5 knew of other F5s and could load balance to other F5s in other availability zones.

Today I'm thinking of the irreversibility of time and leverage.

We use things created by others for maximum opportunity of success. For example we use industry standard companies such as Amazon for our shopping or use open source software in our software, we don't need to create everything by hand.

Given there is an opportunity cost to every decision, and you cannot extract egg from a cake when it's baked, how do you reason what is around to stay? We need to use solid companies and rely on solid technologies for success. We also have mental patterns that either limit us or benefit us.

I'm going to do a social experiment. I'm going to write what I'm thinking of and doing on this comment page over the next few days.

It's your responsibility to try coordinate with me and I'll try coordinate with you.

This can be a low tech recursive coordination portfolio organising space.

To start with my coordination with you Mindey, I shall provide my reaction to your thoughts and feelings. Integrating ODOO sounds like hard work. I suspect it shall be difficult and take a lot of time. It's achievable with lots of effort. You would also need to learn how to host ODOO multitenanted so you could run it on your computer for us to use, unless you expect us to set it up too which is probably also uses lots of time and is a lot of work to set up and keep up to date. I want to focus on solutions that are achievable and impressive at the same time.

Today I spent my day working on Quora and someone solved this stackoverflow problem of mine - I managed to get a balanced Cartesian product. I want to load balance loops. When you have a nested loop of three levels, you run the inner loop more than the outer loops. I wrote code to load balance the loops.

I have an idea for adding similar to Erlang concurrency to any number of loops.

I am also thinking of writing a simple interpreter to my M:N scheduler. And using a simple assembler to write code to memory. And execute it. Then add multithreading. I want to be capable of creating new threads and running code in multiple threads. Java and C# did it so why can't I? It's hard but I kind of know what I need to do. I wrote a parser before. I have written assembly before. I need to write some assembly patterns and write code to generate that assembly. I'll need a new thread primitive. And primitives for compareAndSet and loops. I can write basic assembly to do all these things then write a generator of that assembly and parameterise it. It's like Java's template interpreter.

It won't be portable but it would be good

// Everyone announces what they ARE doing.

This requires "status updates" or "projects as statuses", those streams be public.

I have something emergent in my head about this (publishing protocol), and it is related to as-of-yet-unpublished-yet-viewable-with-this link vision for Odoo. I'm thinking of that Python ERP as a P2P publishing system. It currently is not P2P, but it has a good base. For example, it has eBay, banking and other integrations, that allows it to be a central assets management place for businesses. I don't see why it couldn't shine as a central place to manage social publishing, and inter-operation between thinking spaces like Infinity (I'll have a chance to finalize the API for that), LessWrong, Halfbakery, and other similar places. Publishing follow up streams of events about projects would then be quite easy too.