2016 Talk 2: Jason

[rrnewton]

By "a lot more square brackets than we're used to", did you mean "curly braces"? What was that alluding to?

[david-christiansen]

Why does (member x d o) not have an RHS in the conde? Or rather, why a different number of RHSes?

[david-christiansen]

What is the "diff2" style of constraint? Perhaps the audience will know.

[david-christiansen]

I don't recall you pointing out that _.0 is the syntax for fresh names. That might be good, if you didn't.

[david-christiansen]

The "many languages" slide is very nice.

[ccshan]

Make the "Macros . Functions" slide(s) fill the screen, so that it's easier to see the punctuation change in the middle.

[david-christiansen]

What is "CL reducer/synthesizer"?

[david-christiansen]

The fact that member returns the whole list might be non-obvious to non-Lispers.

[david-christiansen]

the ==constraint slide is nice, but you hopped through them faster than I could read the text and the code together.

[ccshan]

Love your antiunification animation, but can you make it zoom in at the same time so that the resulting 3-in-1 definition fills the screen?

[david-christiansen]

The make-invalid? implementation slide doesn't have text like the ==constraint slide did. Consistency would be nice, I think that without text is better.

[ccshan]

The rule that your slides should contain at most 5 bullet points (total across all pages) is now active.

[david-christiansen]

Is !- the same as the turnstile symbol?

[samth]

• What is the point of this talk? Apparently from the paper it seems to be about adding CLP to microkanren.
• Are you really claiming that this is more efficient than Prolog?
• Don't say that this is different than traditional LP without being very sure.
• How much do you know about what people are used to?
• Fewer transitions, both that remove text or just add it.
• So much code that it's hard to follow. What do you want the audience to take away?
• You don't have to show your system producing ugly output.

[vollmerm]

This was an entertaining talk, but I'm still not sure what your main contribution is. Maybe you could signal when you transition from the "background of minikanren" part of the talk to the "new stuff" part.

[jsiek]

Too much use of incremental builds in the slides, that is, text popping up and/or going away.

[sabry]

What you are really doing seems to be "playing with miniKanren" and your talk just seems to wander around showing interesting little things. But the talk really lacks a crisp motivation and contribution. In answers to questions and "off script" you said some interesting things about combining constraint systems over different domains etc and having a lightweight way of experiment with customized solvers that could make a crisp motivation and structure to the talk.