Very interesting project
though everything here sounds pretty much abstract, and in general it looks like the project lacks focus - meaning it needs to define some specific direction (maybe several ones) and most importantly concrete goals set with their respective timeframes.
I am a bit lost reading about all those capabilities... trying to understand what specific targets this powerfull framework will have. Will play more with website to see if it clears up things.
The framework was not designed specially for blockchain originally, so it's focus is very wide, from hardware drivers ( usb, hid,audio, vesa), asynchronous events, management of different types of network protocol, object references, but it has very pragmatic application, and all the code already works, but it's made to be a modular framework, rather than monolithic application, so necessarily it's a bit abstract
The goal of the framework is to build a solid, interoperable, light weight abstraction layer to build application on, and I have different practical cases in mind.
In a way i kept this project out of the commercial pipeline to have the time to advance on the code at baby step looking the best way to implement each thing with network protocol json/rpc , threading and asynchronous io in mind, and to avoid to have to rush the development of certain thing to release specific application with a well defined dead line, but lots of it is already ready and tested now
For the blockchain specific things, there is already large part of what is needed to make a client for any blockchain based on bitcore , using any hash algo that is compiled in or in a module, with staking with pos 1/2/3 and is interoperable with bitcore based wallet, with compatible rpc interface.
There is not much documentation online, only short explanation on the rough idea, but I will write more detailed documentation latter.
I have just a small question... You tell you will provide universal binaries to be used on practically any system and architecture. I wonder which format those binaries will be based on? It cannot be just assembler/machine code, as this cannot be used both on x86 and ARM, and even I doubt the same format could be used on x86 machines running Windows and Linux respectively. Will that be .dll, .so, or something proprietary?
I say only on different system, the binaries are position independant dynamically linked modules, which contain machine code, they can be run only on same cpu architecture, but they dont depend on the compiler runtime or the host system, all the module are in simple C , it should not be hard to compile for different cpu architectures, including arm, I have plan for the raspberry pi , but it will need a version of the binaries for each cpu architectures, for the moment it's 32bits x86, with sse for the raytracing, it can run on any system running on x86 /x64 with x86 emulation, other binaries need to be made for arm, or for native 64 bits intel.
However I would like to have some test with udp implementation of certain parts of the protocol enough to make useable client, before to get into arm.
The same binaries can already be compiled on windows with visual studio and used on linux, or compiled on linux with gcc and used on windows, it already works and is testable from the git. I already tested this with micro kernel architecture so can safety say the dependencies are "in check".
Most of the binaries running on the ubuntu 64 server currently have been compiled with visual studio 2013 on windows, including the raytracing with the sse code with sse C intrinsic.
It is home made ABI, there is zero documentation on the inner format, but the code source to create them out of 32 bit dll & so file is on the git ( https://github.com/iadix/purenode/tree/master/mod_maker
) , and the loader too ( https://github.com/iadix/purenode/blob/master/libcon/tpo_mod.c
I should release a new version on the git in the next days, and the new block explorer, after i will improve the 3d demo page, and then ill see after in 2 weeks how everything goes with the ico préparation if there are more needs for the site and web things, then I will write more documentation on the memory model, references tree, json rpc, modules, and the blockchain specific modules & applications with more html5 examples, either for the website or for the node built in application like block explorer & secure wallet, and/or with the raytracing.