The expectations for research and the expectations for engineering are different, and I suspect that's where a lot of this comes from. We've kept our burn insanely low (we still have a full startup's worth of runway left), because we knew we didn't have a handle on exactly what the problems were or exactly how to solve them. We're not naive enough to think we can waltz in and do something very smart people have been trying to do for 50 years - we needed to take our time and explore the space. We've tried to communicate that in several ways: always talking about experiments, posting bibliographies, doing a few talks on the crazy ideas we tried, etc. All in all, I'm really proud of what we managed to do in two years - it's easily several PhDs worth of work (our full bibliography is over 300 papers!).

We've had a rationale up for a while [1] and our first post on incidental complexity also went over a lot of that [2]. But given the above it's certainly not as explicit as you're wanting, because until very recently we hadn't figured out the exact what or how. The plan for the fall release is to present a real system for people to use as well as a ton of content to lay all of that out in detail. We'll be drawing lines between all the various things we've done and all the papers we've read to the real working system - tradeoffs and all. We're actually really excited about it - it's been tough figuring out how all this is meant to fit together and we have the opportunity to present things in a very exploratory way, where people can dig as deep as they want (there'll be weeks worth of stuff). Once we pull a first draft of that together, I'd love to send it to you to get your thoughts.

[1]: https://github.com/witheve/eve-experiments/blob/master/desig...

[2]: http://incidentalcomplexity.com/2014/10/16/retrospective/