It’s Monday after flying back from St. Louis yesterday, I had jury duty, and Strange Loop ended Saturday afternoon. Which is to say, I am tired. But! I think I’ll feel better once I’ve translated my notes to some thoughts, and I’ll start with PWL (since I also have Strange Loop notes!). We’ll see if those other post(s) happen, TBD.
I was worried that PWL would be over my head, or that I would feel just oh so lost listening to these very smart people talk about math and proofs and things, and while I could physically feel my brain (does this happen to other people??) work and expand or something, I was very pleased that dare I say, I might’ve enjoyed PWLConf more than the two days of Strange Loop. Perhaps I’ve gotten better (too good?) at metering my energy over the course of the two days of SL, but PWL had banging talks consistently, I found them all interesting, and it was just great!
PWL is single-track, so no decision-making needed (contrast to SL), and starts at a generous “almost 10am” YES.
The first talk was On the Expressive Power of Programming Languages with Shriram Krishnamurthi, presenting a paper he read in grad school (at PWL, people are generally presenting papers other than their own, although this was not 100% true at PWLConf. If this idea piques your interest, I suggest checking out the main PWL site).
- Does feature F add “expressive power” to language L?
- If we can rewrite/translate/compile L+F to L, it does not add expressive power
- The “Las Vegas Principle”:
“What happens inside parentheses, stays inside parentheses”The “Las Vegas Principle” according to Krishnamurthi, or, reading Lisp-like languages
- How do we define equality?
- Discussing “observational equivalence” and I have some note here referencing James H Morris, so it must be a lambda calculus thing (ah, another thing into which I have never delved as deep as I like)
- Is there a way in the language of telling them apart. If not, they are equal. If so, then some program may have been relying on that, and you might break things
- (oh look, some note about lambda calculus, but I didn’t write more, lol)
Loved this talk. It makes me wish I was one of Krishnamurthi’s students! The interesting bit of this talk that I can take (and actually, now that I think about it, I ran into something similar not un-recently in a work context with discussing changing comparison behavior) is that adding expressiveness can break previous comparisons.
In next talk, Giulia Fanti presented her own work on anonymity + Bitcoin (spoilers, it’s not anonymous!). This talk was especially well done because I am predisposed to hear anything about the “blockchain” and immediately become bored and want to listen in a different direction. BUT Fanti got my attention because I am doing more work considering graphs lately, and blockchain is loootttts of graphs.
She went over the history of the protocols (trickle vs diffusion) used, and research into if diffusion actually improved privacy on the block chain (ehhhh not really! pretty easy to de-anonymize).
Then she presented her own research’s proposal, the Dandelion protocol, which is already used on Bitmessage (which is evidently not that popular, but research implemented is cool!). Also, it’s cool because the name is from how the graph looks 🙂
After the lunch break (food trucks! fun), the next talk was again someone’s own research, this time on program analysis with Karim Ali, which I found interesting since I’m not super familiar with this other than “I have found ASTs useful.”
Unfortunately (or fortunately for furthering research?), turns out program analysis isn’t very popular/people don’t want to use it because it scales poorly, has not-gread usability, and precision is either too noisy, or you miss something important.
In my notes, I got a bit lost because somewhere near the beginning, a paper was referenced, but I couldn’t catch if the talk was on that p
- Averroes – partial program analysis (more scalable! much win!) – this is from Ali’s lab
- SPDS – Synchronized Pushdown Systems, program analysis research
- CogniCrypt – an Eclipse plug-in to enforce good crypto practices, uses SPDS!
- A paper I want to read: Social Influences on Secure Development Tool Adoption (this sounds very interesting)
- “Cheetah” – an Eclipse plugin analyzing Android applications
Generally enjoyed the talk, but I was confused about if it was a paper talk, or a “this is my research” talk, so it might’ve been a tech talk with lots of research mentioned.
I took a break for the next session, because I wanted to be well rested for Heidi Howard, who was presenting on distributed consensus. Howard’s thesis was covered in the popular the morning paper in multiple parts (I have not read this yet).
I don’t have a lot of notes from this (late in the day! waning!) but I really enjoyed Howard’s walk-through, which made a complex topic feel accessible. But I still think distributed consensus is super cool/very interesting.
Finally, the last talk of the day was Star Simpson talking about a 2005 paper on how the in-car voice impacts driver behavior. It’s super interesting, and I love that she closed out the conference (thus the opening and the closing both did this!) doing the presenting-someone-else’s-paper pattern.
The paper she discussed was thought provoking – the tone of voice had a measurable impact on driver behavior when drivers had differing (upbeat vs agitated) states of mind. This makes me think of a lot more things though: thoughts on psychology and how there are lots of things we do as humans that we don’t think about (changing our tone when dealing with an upset person being a topical example) where we change in reaction to a given context.
And that was Papers We Love Conf 2019! It was really interesting, and very well curated and organized (from the attendee perspective). Very grateful to the speakers and organizers who made it possible! Now I have a lot of reading (and writing!) to do …