Has there been any work on the development environment for AT? The spec at CIYAM seems very low-level.
Is a crypto API going to be added to it? Doing crypto using the AT instructions would be very expensive.
How does the AT code interact with reward assignment? Vitalik Buterin thought up a mining contract where the terms of mining would be very enticing, and would lock participants' coin into the contract using a heavy penalty once the mining pool controlled more than 50% of the hashing power, giving the controller of the contract the power to double spend. Such a malicious arrangement could be simplified if AT code can automatically set reward assignment.
It would be great to see the source code for the AT branch of burstcoin. Is it available anywhere?
Yeah, it is indeed assembly code that will be working initially, just yesterday started working on trying to get a simple compiler written.. anyone want to help, let me know.
That being said, assembly isn't impossible to write on your own. and you get control of the program and can write it very efficiently, which might be important to some when posting by the instruction.
Dev had plans to figure out reward assignments over weekend as previously every instruction was being considered equal. You want to discuss further with him, he might appreciate it.
But it's important from a security standpoint that ATs are written in assembly in the core code because that makes it easier to catch bugs. If we build a higher level language on that can compiler down to assembly, that's ideal and intentional.
One other thing, vbecas and probably me (I'd want to be paid), maybe others, would be willing write ATs/help people learn if they need help getting started.