Initial pass over presentation

[rolyp]

Skeleton of talk for practice run at PLUG, Tue 11 Jan.

Opening slide

• [x] title; authors + photos

Problem

• [x] understanding visual outputs means understanding what visual elements represent
  • [x] examples: various typical (= opaque) charts (see HPI talk)
• [x] can we do better?
  • [x] open source/open data help provide the raw materials but no insights into how things are related
  • what we would like is computational artifacts able to reveal their relationship to data interactively, on demand (without incurring programmer overhead)
• [x] two specific problems we would like to make progress towards; first is linking outputs to inputs in a fine-grained way
  • [x] author just writes regular functional viz code -- bar chart example
  • [x] "round-tripping" properties important for trust/correctness (here: sufficiency)
• [x] second is linking outputs to other outputs in a fine-grained way (and seeing the relevant data) -- line chart example
  • [x] for the forward analysis here, we care about necessity, not sufficiency

Solution (part 1)

• [x] prior work: Galois slicing provides a way to relate input selections to output selections, with nice round-tripping properties characterised by Galois connections
• [x] Galois connection: best way to (nearly) invert a map that can't be inverted
  • [x] backwards question: what is the (least) input needed for this output?
  • [x] forwards question: what is the (greatest) output that this input suffices for?
• [x] in Galois slicing, a selection of an input/output is a prefix (tree with holes)
  • [x] example from Fig. 16
• [x] however, doesn't support isolating arbitrary components of structured outputs
  • [x] differential slicing helps, but wrong approximation for this problem (example from Fig. 16 -- explain "spine")
• [x] Galois dependencies
  • [x] lift prefix-closure restriction, so "selections" are not just terms with holes; selections are just subsets of the set of paths in a term
  • [x] represent by associating a bit with each path (more generally: elements of boolean algebras) -- give grammar of values
  • [x] less "program-like", but more fine-grained
• [x] forward/backward analyses
  • [x] form/meaning of the judgements
  • also introduce hole as shorthand for .. (relevant to environment merge during backward analysis)
• [x] demo: convolution example

Solution (part 2)

• [x] also want to understand how outputs relate to other outputs
  • [x] nature of "forward" question changes: not what x is sufficient for but what x is necessary for
  • [x] turns out our more general notion of selection can help here too, because it supports negation
• [x] De Morgan dual
  • [x] using negation (complement), can formulate "necessary for" in terms of "sufficient for"
  • [x] think of this as inverting the direction of the GC (by turning meet-preserving into join-preserving and vice versa)
  • more generally: composing a GC with itself
• [x] demo: linking example

Closing slide

• [x] conclusion & future work
  • [x] link to f.luid.org
  • [x] future work: expression provenance

Done/dropped:

• [x] settle on slide solution (Keynote: no; Autodesk SketchBook: yes)
• [x] experiment with mic in bathroom and up close (bathroom: no; up close: yes)
• [x] play around with Keynote with iPad/Apple Pencil (outcome: shite)
• [x] experiment with recording demo (speedup x 4; can embed in Keynote, but not SketchBook)
• [x] check duration of talk: 20+5 min (see "Accepted submission" email of 28 Sep)
• [x] quick pass (unrecorded) to identify main remaining problems/omissions
• [x] define Galois connection using unit/counit formulation rather than isomorphism of hom sets
• [x] remove blue circle around Chiang Main urban area
• [x] drop "spine" explanation
• [x] pass over layers in Goal 2 solution slide
• [x] split De Morgan duals into two slides; use earlier colour scheme for GCs
• [x] reorder slides