Review full draft of technical sections

[spmaniato]

The draft is at a point where the technical sections can be read and reviewed as a whole:

• Preliminaries
• Problem Statement
• Discrete Abstraction
• Formal Task Specification
• High-level Behavior Synthesis
• ROS Implementation
• Experimental Validation

(Note to self: I have too many sections; some are tiny. Maybe combine Synthesis and Implementation.)

I'm well aware that these sections are nowhere near "camera-ready". But I'm hoping to get some feedback sooner rather than later. I've also left the TODO notes in there to give you an idea of what I'm struggling with or am unsure about.

You can view main.pdf online or download it (click the "Raw" button if you don't want the entire repo).

You can give me feedback in any way you prefer:

• Comments on this issue
• Email
• Annotations on the .pdf
• Comments/fixes in the .tex files

Thank you :smiley:

PS. I've assigned to HKG, but I could use feedback from @pschillinger and @dcconner too. Especially as it pertains to #2

[pschillinger]

I read the listed parts, just a few comments:

• III. Problem 1: I suggest \mathcal{S} for "system" instead of s or S (both versions used), since s and S is already used for states in in II.B. Also, in my opinion, \mathcal{S} looks more like system and it would better align with \mathcal{A}, \mathcal{D}, \mathcal{T}, ...
• III. Problem 3, your note: "software implementation" is good. This is a general problem statement and not specific to any realization. Furthermore, the result of a complete synthesis should really be a software implementation (e.g. generated using FlexBE).
• IV.A.: Are actions itself or the (completed) execution of an action possible preconditions?
• IV.B: Similarly, you say the distinction between actions and control modes is purely illustrative. I would argue that this is slightly inconsistent, since control modes rather equal completed actions (of switching to a control mode).
• IV.B.: "outcomes that are of most interest" -> The definition looks like you only consider c and f as outcome. Thus, I would remove or replace the "most".
• V.B.: In IV.A. you allow self-transitions. (3) prohibits them. Do I get this right? In any case, one sentence for clarification could help.
• V.B.: I understand you classify the formulas regarding the type of propositions they are applied to. However, I don't really get the reason for classification since they appear to be very similar (e.g. why is (6) applied to actions, not to primitive functionality like (3)). Btw. what is the difference of the outcome of an activation and the outcome of an action? There can't be an outcome of an action without its prior activation, right?
• V.B.: "system from winning the game" -> I think you didn't use this terminology before in the paper, might not be the best place to use it first. Also, I don't really understand the why the disjunction in (7) prevents anything since a disjunction extends the set of inputs resulting in True. Or is "system" meant to be "environment"?

(skipped rest of V. for now since plane was almost arriving)

• VII. end: autonomy field in message is wrong -> autonomy: [0,0,0,0] (should have same length as outcome list; probably not noticed because all unspecified outcomes get the default autonomy of 0)
• VIII.B.: "locked" (paused) -> "locked" (prevented from returning an outcome), more accurately resembles what locking means

I also made some annotations marking the spots I mentioned above: https://drive.google.com/file/d/0B4uq3dAo3itfS05NMGotc21md1E/view?usp=sharing

[spmaniato]

Thanks for the comments Philipp :smiley:

(I won't reply to each one here. I will address them in the paper directly.)

[spmaniato]

I just pushed (18f76a71212e77b1aa23ccfddbb60a3281662cd7) a full draft of the paper, without TODO notes (main.pdf)

I have incorporated most, if not all, of David's and Philipp's feedback. And I think I finally have more or less consistent notation (also see comment in #2). We are also doing pretty well length-wise (7.5 pages).

All, please take a look and let me know if you have new comments. Thank you :smiley:

PS. The Intro and Conclusion sections are pretty rough drafts. You've been warned :stuck_out_tongue_closed_eyes:

[spmaniato]

This draft has now been reviewed by the authors and a bunch of other people. I'm closing this issue and I'll open a new one once I have a "final draft" (to be followed by "final_draft_v2" and "final_draft_v3_FINAL" :stuck_out_tongue_winking_eye:)