Turing practice talk

[rolyp]

Switch to Keynote instead of iPad.

• [x] Keynote import (with videos)
  • [x] title slide
  • [x] provenance-aware data visualisation
  • [x] Goal 1 (rework as use case – see above)
  • [x] Goal 2 (rework as use case – see above)
  • [x] prior work 1: Galois slicing (drop)
  • [x] prior work 2: Galois connections (rework)
  • [x] prior work 3: Differential slicing (drop)
  • [x] prior work 4: Differential slicing (merge with previous slide and rework)
  • [x] Galois dependencies (rework as “requirements” slide – see above)
  • [x] representing selections
  • [x] forward and backward dependency analysis
  • [x] Demo 1: 1 (simplify, merge into use case 1?)
  • [x] demo 1: 2 (drop)
  • [x] Goal 2 exposition
  • [x] De Morgan duals 1
  • [x] De Morgan duals 2
  • [x] Demo 2 (merge into goal 2 – use to motivate explorable explanations)
  • [x] conclusion (move to beginning))
  • [x] closing slide
• [x] summary slide explaining two use cases
  • [x] programmer just writes regular (functional) code
  • [x] link to data for transparency
  • [x] link outputs to other outputs as comprehension aid (view coordination)
  • [x] convolution example – explorable explanations
• [x] Demo 1 (linked charts)
  • [x] code slide
• [x] Demo 2 (convolution)
  • [x] code slide
• [x] Phil W: why linking important? relevance of meet/join preservation?
• [x] Simon F: say what you need/why prior work doesn’t have it; otherwise kill most of prior work section
• [x] James McK: pare back/drop the goal slides in favour of the videos?
• [x] dry run (26 mins)
• [x] why is this a problem? slide back to the end
• [x] our own programming language
• [x] “baked in” in bold?
• [x] properties we care about -> desiderata
• [x] rehearse from “use cases summary” to “problem spec”
• [x] preliminaries
  • [x] set Keynote to present from laptop screen
  • [x] mic on and nearby

[rolyp]

Jeremy Singer to Everyone (15:00) @James - attending events in even numbers is entirely reasonable :-)

Ornela Dardha to Everyone (15:05) now I can… :)

Thomas Koehler to Everyone (15:24) is this comparable to checking whether gradients are equal to 0 with something like automatic differentiation?

James McKinna to Me (Direct Message) (15:34) Roly, talk looks very interesting, but IO realise I don't have the bandwidth to take it on right now. best of luck with the fina; presentation, but you have way too much graphical and demo material... POPLcan be very unforgiving! And, my personal reaction is to find the mixed tablet handwritten+machine generated content to be very hard to read-at-a-glance... just my 0.02c worth

I'd make the conclusion slide the opening slide!

James McKinna to Me (Direct Message) (15:37) And minimise the 'exploratory' discusison of the scenarios in favour of the two videos? And maybe just pick one example to demo, and then follow-up with a list of what else you might show.

Anyway, I gotta go, sorry. Good luck!

Ohad Kammar to Everyone (15:44) It is a choice though! Because the fact that you're multiplying by 0 might be due to a value in your dataset!

Ohad Kammar to Everyone (15:50) OK, thanks and good luck! I think my feedback has been said by others. See you next week!

Simon Gay to Everyone (15:51) Thanks for the talk, I have to go to prepare for the next meeting.

Ornela Dardha to Everyone (15:56) ok, I’ve got to run! good luck Roly! :)