Monday, June 23, 2014

Another spectacular PLDI

PLDI this year was amazing. I was surprised at the quality of the talks in general, and I especially enjoyed the "hallway track" conversations that surrounded them. Edinburgh was a great setting, too—there was plenty to explore before and after the conference.

Unsurprisingly, I was left with lots of opinions about the details of organizing a conference, from the significant to the mundane:
- I'm really happy that PLDI has gotten on the "lightning session" (teaser talk) bandwagon. Especially for multi-tracked conferences, it's really critical to judge which talks are likely to appeal.
- I'm less enthusiastic about poster sessions. I like having the opportunity to chat in more depth with presenters, but I don't think posters themselves are particularly useful—the in-person conversation is way more important, people generally don't seem to look at the visual aids much, and they take a lot of time to make in proportion to how useful they are. The sticker-based whisky sweepstakes were a nice touch at PLDI, though.
- As I think others have posted here, I wished many times for places to sit down during lunch. Stand-up lunches are space-efficient but exhausting.

I've also posed a blog-form summary of the paper I presented at PLDI, which I'll avoid duplicating here.

Saturday, June 21, 2014

Back from Edinburgh

Attending PLDI has been a great experience for me. This is the first time I attended and presented at PLDI. I was anxious the days leading to the conference, mostly worried about my talk.  It helped that my talk was in the first session on the first day. My talk on "Test-Driven Repair of Data Races in Structured Parallel Programs" was well attended and I also got some interesting questions. I also had a poster on the same topic during which I received some very valuable suggestions and feedback. 

The general quality of the papers and presentations at PLDI were great. I liked the sessions on Parallelism, Scripting Languages, Synthesis and Races. I felt that the scheduling of the sessions could be better -- for instance having Parallelism 1 in parallel with Memory models didn't seem ideal.

I also got an opportunity to interact with Professors and leading researchers in programming languages. I sincerely thank SIGPLAN/NSF for the travel grant and My advisor Professor. Vivek Sarkar for making this trip possible.

Also my first PLDI

I really have to say I enjoyed PLDI this year. Besides extremely interesting talks, making connections to researchers is something I find invaluable. Also, presenting my poster on "Finding Minimum Type Error Sources" was a great experience. In the same spirit, I really plan to visit Edinburgh again. The city is really something special. I am looking forward to next PLDI!


Thursday, June 19, 2014

My first PLDI

I thoroughly enjoyed my first PLDI. I found the list of accepted papers very interesting, I really enjoyed the Parallelism tracks. The one minute teasers was an excellent idea, it helped me choose which talks I wanted to achieve. The teaser by Martin Rinard was especially memorable.

The Future of PLDI session was also very enlightening. I enjoyed Jan and Sriram's impersonation of each other. I like the idea of having more dedicated reviewers per topic and the idea of the two-phase review process. I would also like to vote for automatic upgrade of the PLDI papers to Journals (with the authors having the option of submitting an extended version which is not bound by the page limits of PLDI).

My talks at the associated workshops (X10 and ARRAY) went very well too, got lots of interesting questions and positive feedback. I also had the opportunity to network and discuss with researchers in similar fields.

I'm also glad that I got to go to Edinburgh and experience the city. It is charming beautiful. After the conference, I also made a small trip around the Highlands - Isle of Skye, Loch Ness, Loch Lomond, etc. It is so green and peaceful. Definitely plan to visit Scotland again.

The poster session was very helpful, I got lots of interested listeners and useful feedback. I am very grateful to NSF-SIGPLAN for the travel support that I received. Looking forward to attending PLDI 2015.

Wednesday, June 18, 2014

Post PLDI

ISMM was amazing!!  It was great listening to what others were working on in terms of designing better garbage collectors, etc.  Even though my memory management background is not that strong, I was able to get the gist of many of the presentations.  It was very insightful listening to the key ideas and tricks that others presented.

This was also the first time that I gave a talk at a conference (as well as my first conference!).  It was definitely nerve-racking at first, but the anxiety went away soon afterwards.  I would definitely look forward to attending future conferences!

I'm also glad that I got to go to Edinburgh and experience the city.  The architecture was very pretty and the food was delicious.  I'd like to thank ACM, NSF, and Professor Zhang for giving me this opportunity!

Tuesday, June 17, 2014

Attending PLDI was an awesome experience

Finally I was able to attend PLDI@Edinburgh after much dilemma with my VISA. The experience was pretty good. 
I attended almost half of the talks and those were really interesting. The keynote speeches were excellent. 
I had a paper titled "Accurate Application Progress Analysis for Large-Scale Parallel Debugging" and presented a poster of the same. Several attendees were interested in my work and had several interesting discussions during the poster session. My paper was on "automated" root cause analysis for large scale parallel applications which run on super computers. During my discussions with other researchers I realized how similar techniques can be extended to work in other environments as well.
I also met with many fellow students from other universities, go to know about their research experiences. 
All these happened because of the NSF-SIGPLAN travel support that I received. I am really thankful to them for giving students the opportunity to attend PLDI. I hope this experience will help me a lot in future.

Post-PLDI Blog

I had a great time at the meeting this year.  The poster session was a great way to get to meet people, and there were quite a few interesting talks.  I was particularly fascinated by Krishna Palem's keynote talk on inexact computing, and the rest of the talks on approximate computing.  It is a new thought to me, and seems to have some really neat implications for all sorts of problems (e.g. scientific computing, and time-critical decision making problems like weather forecasting or stock trading).  While the notion that we can't rule out certain inputs yielding catastrophically incorrect results (as demonstrated in Eric Schkufza's talk "Stochastic Optimization of Floating-Point Programs using Tunable Precision") is somewhat unsettling, I think that may only be because we are so used to treating computers as fully deterministic machines.  However, we certainly make quite a few decisions based on computations we do not fully understand (particularly in the case of human intuition)!

I also really enjoyed the "Future of PLDI" discussion.  I resonated with a few of the things said, including the idea of making some optimizations on the reviewing process.  In my opinion, having each reviewer handle 20 papers is too much; it cannot allow enough time for reviewers to obtain an in-depth understanding of each paper.  Reducing the number of papers assigned to each reviewer and reducing redundancy in reviewing resubmitted papers could go hand-in hand.  One idea was to allow authors to submit with a list of conferences they would be willing to publish to.  Since authors are likely to re-submit the same work to lower-tier conferences, the reuse of reviewers would save reviewer-time.  Another idea I liked was to pre-publish on a site like arXiv and encourage peer feedback prior to publication.  This would improve the quality of submissions, likely streamlining the review process.

All in all, attending PLDI gave me a chance to meet fellow researchers, and get better acquainted with the community and research at large.  Thanks for the opportunity!

Jake Brock

Monday, June 16, 2014

After PLDI

This is Xin, writing from Atlanta, four days after PLDI'14. As I mentioned in my previous post, I expected this PLDI to be a wonderful experience, and it didn't fail my expectation. I especially thank NSF for its generous support for my trip.

I actually had quite a busy time before and during the PLDI. I had one paper and one poster to present. Plus, I need to help work on the slides of another paper of mine, which is presented by another coauthor. Basically, I spent the whole couple of weeks before PLDI working with PowerPoint and rehearsing my talk.

However, all these efforts didn't end up in vain. I had a great time during the poster session and presenting the paper. I had a good time chatting with people and received interesting feedback. It felt especially good when others came to you and expressed how much they liked your work. During the poster session, I even had a chance to describe my work to Neil Jones, the receiver of Programming Languages Achievement Award this year.

Having felt so good about this year's PLDI, it makes me look forward to future PLDIs.

-Xin

Post-PLDI Post

This year's PLDI was a pleasure to attend. I had an excellent time meeting other researchers, hearing about their work, and sharing my own. The lunches and dinners were set up to allow easy mingling, which was nice. (I'm glad weren't seated at tables for lunch.)

The poster session was excellent. It was a good way to engage people in a 1-on-1 format, which can be a much more focused way to share your work. I wish I had time to take a look at other students' posters, instead of being glued to my own!

I also gave a talk. It was my first major conference talk, and I was terribly nervous as I gave it, but it seemed to go well despite that. The questions after the talk indicated an insightful and attentive audience.

My only complaint is that the ballroom projector wobbled whenever anyone was near it. Given the quality of the conference in every other regard, it seemed a little silly to have a bobbling projector :-).

Thanks to the NSF for funding.

Hope to see everyone I met again in the future,
Justin

Sunday, June 15, 2014

This year Scotch, next year Coffee

Having attended the main conference, as well as the co-located events, I have to say that this was one of the best times I've had at a conference yet. I especially enjoyed seeing the broad range of work in the APPROX workshop and chatting with people about research throughout the conference. It makes me very excited to be in this field. I'm looking forward to PLDI next year. The coffee situation was a little weak this year, so I'm stoked to head off to Portland next year! I'm also hoping we'll see more ladies in Portland -- it will be co-located with Grad Cohort, so maybe we can get more people attending both events!

PLDI Wrap-up

This was a great year for PLDI. One of the themes of the conference this year seemed to be a sort of "ruthless pragmatism" which made me quite happy. I enjoyed most of the talks I went to and met a lot of interesting folks. I too thought that the one minute talks were a great idea. They gave a flavor of the work being done across the whole conference without requiring the listener to attend every full-length presentation.

Our work on Verification Modulo Versions was well-received. Francesco gave a great presentation and I had several discussions with people who enjoyed the paper/talk.

The SRC poster session and talk also went well. I had an especially interesting discussion about the relationship of what we're doing to some concepts in machine learning, which could be a very interesting future research direction. I managed to finish my talk on time and got some excellent questions from the audience. I really liked hearing the other SRC talks and getting to talk with some of the other students about their work--some really interesting ideas in there!

Hope to be back for PLDI '15 in lovely Portland.

What happened at my first PLDI?

My overall impression of PLDI14 was really good. I found the list of accepted papers very interesting and I was able to talk to many people about my research.
The paper I presented on my language FAST received very good comments and many people showed interest in the topic.
Attending and presenting at PLOOC, the workshop on PL technologies in MOOCs, was also very helpful.
I loved the 1 minute talks at the beginning of the day. They were very helpful in choosing which talks to attend (or not attend).

As many pointed out few things fell a bit short:

  • The parallel sessions were poorly organized. The two main examples being Atomicity and Parallelism happening at the same time. The same for Verification and Static Analysis 1.
  • The second room for the "B-track" wasn't adequate. It felt like there were business class and economy class papers. The first room was a theatre, the second one was awful with bad audio, a shaky projector and not enough chairs.
  • Lunch: inadequate portions but most important it was not possible to sit. For many shy students (especially first year PhD) it's not easy to go and talk to many professors. Lunch is usually one of those occasions where it is enough to sit at the table to start a discussion. This wasn't an opportunity this time.
I'm overall very satisfied and these few problems do not affect how useful the conference was for me, especially in terms of networking and visibility. Looking forward to attend PLDI15.

The conclusion of PLDI'14

As many attendees, I return right after the main conference, regretting a bit for not staying one or two days more to explore UK. Nonetheless, the conference experience is great for me. I enjoy the city of Edinburgh a lot. It's such a nice and pretty city! And for that, I'm grateful for NSF's and ACM's support in this trip. I would have never thought of visiting Edinburgh even in my wildest dream.

I have to admit that I do not talk to as many professors as I should/am expected to. Most of the time I find myself talking to other students more. This needs to be changed in the future. But on the bright side, both students and professors whom I talk to are genuinely interested in my Big Data-related projects. That simply gives me much-needed encouragement.

This is also my first time to present a poster. Even though I had some practice at home, I still find it difficult to present my work in the event. Because of the loudness, the crowd or just me being over-nervous? I'm not sure. Luckily the later talk (even though is short - 9 mins) goes well.

The walking tour is absolutely awesome for knowing the history of the city a little bit (despite the main theme is all about horrible and creepy things. On a side note: I still find it amusing to see a tent while walking in the cemetery :) )

I also learn a lot from attending the talks: things to do and not to do when giving a talk. 

This trip is great and I really want to attend the next one (here in Portland, OR)

Best,

Khanh Nguyen

P/S: I do hope that the next conference will not serve us lunch with such small portions :D




Thursday, June 12, 2014

Post-PLDI

A villanelle

It was a joy to attend and present
our research to renowned scientists.
(Thanks to the ACM and NSF.)

After rehearsing time and time again,
the time had come to bring Shapes to the show.
It was a joy to attend and present.

A teaser, poster show, and main event,
the third run by coauthor Fabian M.
(Thanks to the ACM and NSF.)

All went well 'till a man in the back said,
"I invented this stuff score years ago!"
It was a joy to attend and present.

Silenced but not appeased the old man left
We carefully considered his comments.
(Thanks to the ACM and NSF.)

Typeclasses could have reined F-Bounds; instead,
industry shaped a fix over the years.
It was a joy to attend and present.
(Thanks to the ACM and NSF.)

Monday, June 9, 2014

PLDI

My name is Michael Carbin and this will be PLDI number four for me.  I am an nth-year graduate student at MIT that will soon be graduated and then soon thereafter be an assistant professor at a yet-to-be named institution.  

This has been a hectic year for myself with spending the last four months on the road interviewing for faculty positions, however I am looking forward to catching up with old friends at PLDI and also presenting at APPROX (the first SIGPLAN workshop on Probabilistic and Approximate Computing) this Friday.

APPROX is not to be missed!

1st PLDI

Hi!

I am Jesse Huang, an undergraduate at Rutgers University and working with Professor Zhang. I am participating in the International Symposium on Memory Management (ISMM), which is co-located with PLDI.  We are exploring how to use atomic operations extensively for computation rather than communication on GPUs.  I will be presenting our discoveries on Thursday.

This will be the first conference that I will attend!  I'm pretty excited, since this will be a new experience for me.  I'd like to thank NSF for helping fund me to attend this conference!

Pre-PLDI

I am happy to be returning to my second PLDI, and in such a beautiful setting!  I took a PLDI course since my last PLDI conference, so I am excited to hear about some of the research going on (now that I have a better grasp of the basics).  The presentations I am looking forward to include: "A Framework for Enhancing Data Reuse via Associative Reordering (Stock, et al.)", "Laws of Concurrent Programming (Hoare)", "Resugaring: Lifting Evaluation Sequences through Syntactic Sugar (Pombrio and Krishnamurthi)".

In addition, I will be presenting my poster on Monday: "Fair Partition-Sharing for Multicore Caches", and my fellow Rochester students' paper at MSPC on Friday: "Affinity-Based Hash Tables".  I look forward to getting to hear all about the latest research, and meet a few people along the way!

Jake Brock

Sunday, June 8, 2014

PLDI First Timer

Hi everyone,

I just finished my third year at University of Massachusetts and first year in the Programming Languages and Systems lab. This is my first PLDI. At the SRC I will be presenting my work on SurveyMan, a programming language and runtime system for designing, debugging, and deploying scientific surveys at scale. I keep a research blog, where you can read more about current directions of the research.

I'm grateful to the NSF for providing funding to allow me to be here and am looking forward to meeting everyone!

Emma Tosch

Saturday, June 7, 2014

On the way to PLDI

I'm writing this from the Amsterdam airport, after the first leg of my trek from Seattle to Edinburgh. (Here's an obligatory conference travel selfie, complete with poster, to prove it.) I'm exhausted from the transatlantic flight but excited to be going back to PLDI, certainly one of my favorite conferences all year.

I'm cobbling together a talk for the conference itself (paper: Expressing and Verifying Probabilistic Assertions) as well as a slot in APPROX, a new-this-year workshop on approximate and probabilistic computing. I'm both excited and nervous for both.

Since conference talks are on my mind, I wrote a longish blog post about lessons I've learned so far in grad school about preparing them. Rather than clog up this page with that text, I posted it to my own blog.

See you in Edinburgh!

Thursday, June 5, 2014

Hello from Providence, RI

Hi all,

I'm a second year PhD student at Brown University, advised by Shriram Krishnamurthi. This will be my first time attending PLDI, as well as my first time visiting Edinburgh. People keep going "ooh, Edinburgh!" and then suggesting sites to visit, so I look forward to the trip. My thanks to the NSF for helping fund me to attend PLDI.

I'm generally interested in programming languages, with a special interest in syntactic sugar. I'll be presenting in the main conference (Tuesday, Olio, 15:45):

Resugaring: Lifting Evaluation Sequences through Syntactic Sugar
Syntactic sugar is pervasive in language technology. It is used to
shrink the size of a core language; to define domain-specific
languages; and even to let programmers extend their language.
Unfortunately, syntactic sugar is eliminated by transformation, so the
resulting programs become unfamiliar to authors. Thus, it comes
at a price: it obscures the relationship between the user’s source
program and the program being evaluated. 
We address this problem by showing how to compute reduction
steps in terms of the surface syntax. Each step in the surface
language emulates one or more steps in the core language. The
computed steps hide the transformation, thus maintaining the
abstraction provided by the surface language. We make these statements
about emulation and abstraction precise, prove that they hold in our
formalism, and verify part of the system in Coq. We have
implemented this work and applied it to three very different languages.

I look forward to meeting many of you at PLDI!
Justin Pombrio

Tuesday, June 3, 2014

My Third PLDI

Hi all,

I am Xin Zhang, a third year Ph.D from Georgia Tech. I have been fortunate to attend PLDI since my first year with the generous support from my advisor Mayur Naik, SIGPLAN and NSF.

PLDI has always been a major festival of the year for me. Not only is it a chance to hear talks on top-quality publications of our field, but it is also a rare chance to meet people in the community. Every time after PLDI, I always got back with fresh ideas and energies to work on research. I am confident I will feel the same after this PLDI.

Our group has two papers accepted at PLDI this time, and I am one of the authors for both papers.

"On abstraction refinement for program analyses in Datalog" proposes a novel counterexample-guided refinement (CEGAR) based approach via partial MAXSAT which automatically searches efficient abstractions for a given program analysis in Datalog. For a given query like a aliasing check (v!=u), our approach will either find a cheapest abstraction which proves this query or conclude that there is no abstraction in the space that can prove the query. The key observation here is that there is a natural connection between Datalog and SAT formula. Each grounded Datalog rule is basically a Horn clause. By encoding the derivations in Datalog as the hard constraints in MAXSAT, our approach successfully avoids all the counterexamples. By encoding the cost of the abstractions as weights of soft constraints in MAXSAT, our approach effectively chooses the cheapest abstraction while avoiding all the unviable abstractions.

"Hybrid top-down and bottom-up interprocedural analysis" introduces a novel interprocedural analysis framework which combines conventional top-down analysis and bottom-up analysis, and outperforms both empirically.

I am looking forward to seeing you guys at PLDI!

-Xin

Monday, June 2, 2014

Pre-PLDI

Hi all,

I'm currently finishing up my third year in the PhD. program at Washington University in St. Louis and will be attending my first PLDI.  I will be presenting work done by my co-authors and I at the co-located workshop MSPC.  Our paper, "Trash in Cache: Detecting Eternally Silent Stores" discusses a technique to track liveness of objects while they are in cache, allowing us to clear dirty bits and avoid write-backs of any dead data in cache.  I will also be presenting a poster based on this work Monday night during the SRC.  I look forward to answering any questions and receiving feedback from everyone.

I am very excited to attend PLDI thanks in part to NSF funding through ACM SigPlan.  I look forward to hearing as many talks as possible.  I also look forward to discussing research and life with as many of you as possible.  See you all there and safe travels.

Jon Shidal

First time at PLDI

Hi All,

My name is Zvonimir Pavlinovic. I am a first year PhD student at NYU, advised by Thomas Wies. I am generally interested in programming languages, with a current focus on automatic localization and repair of type errors. I will be participating at PLDI Student Research Competition with my work on how can we leverage SMT solvers to enable compilers localize type errors in functional programming languages like OCaml.

This is the first time I am attending PLDI and also visiting Edinburgh. I am quite excited since there is a bunch of really interesting papers appearing at the conference, and Edinburgh seems to be a really cool place as well. 

See you all there!




Sunday, June 1, 2014

Hello from Santa Barbara

I'm Vineeth Kashyap, a PhD student working with Prof. Ben Hardekopf at UC Santa Barbara. I am excited to be attending my third PLDI, as well as traveling to Edinburgh. The talks at PLDI have always been of very high standards, and the works presented always peek my interest, with a bit of PL theory and some impressive implementation.

I have been working on a novel JavaScript abstract interpreter framework (which I will be presenting at SRC), and building a variety of client analyses on top of it. I am looking forward to get some inputs from PL experts at the conference, and see if anybody would find our analysis code useful for their research. I also managed to read some papers from the PLDI program, and it would be nice to talk to the authors in person. There are some papers directly relevant to my current research projects: "Speciaization Slicing", "Introspective Analysis: Context-Sensitivity, Across the Board", and "Selective Context-Sensitivity Guided by Impact Pre-Analysis". Apart from all the static analysis related talks that I am interested in, I found some really cool papers by fellow SRC attendants: "Fast: a Transducer-Based Language for Tree Manipulation" and "Getting F-Bounded Polymorphism into Shape". I hope to write about these in my post-PLDI blog.

I am also shopping around for postdoc positions and industrial research jobs, so hopefully PLDI will be a fruitful place to network. See you all there!