Thursday, November 7, 2019

SOSP19 Day 1 wrap up

It was only 3 sessions into day 1, and my brain was fried.
Conferences are tiring because you are exposed to so many new ideas in a short time. It was clear I would not be able to pay attention to the papers in the last session, so I skipped that session (the privacy session which included the following three papers) to go for a walk at the golf park behind the conference center.



After the privacy session, there was a poster session and reception from 5-7:30pm. The poster session was nice for asking authors questions about the papers and having more in-depth conversation.

A student had told me he doesn't know how to start conversations with other conference attendees. I told him "That's easy.. Just ask them about what they are working on these days." A better way to start deeper conversations is to listen to the paper presentations, and have genuine questions about future work, or some extension and connection, and go discuss with them at coffee breaks, lunch, or poster session.

In the free-roaming poster session and reception, I had a chance to meet many colleagues and catch up on what they are working these days. When they returned the question, I had to talk for 3-5 minutes about what I am working on these days. I found that my "elevator pitch" got better and better as I had to answer this question many times.

I am a shy person, but at conferences my curiosity works in my favor, and I approach people to learn about their current work, and what they think of this paper versus that paper. I really enjoy talking to fellow researchers, each of whom is an expert in a small part of a big field. We may have different opinions on things, they may not like the papers/ideas I like, but I get to learn about their perspectives and file them in my brain without having to agree or disagree with them for now.

General impressions about SOSP 

SOSP is single track, so 500+ people were in the same big conference room for the sessions. The first half of the room had tables, and the second half just chairs. If you sat at a table row, you can rest your laptop on the table and type comfortably. I sat at the very front row and took notes. Interestingly, there is little contention for the front rows. Another advantage of sitting at the front row is that I am not distracted by seeing other audience members checking Facebook, Twitter, and mail on their laptops.

(Rant: This is my pet peeve. This drives me nuts. What is the point of flying over a long distance, driving at least two more hours from the airport to come to the conference, and check mail and social media all day long? You disrupted your whole week to travel to this conference and now you are not "at the conference". This I will never understand. Be here now!)

To comment on the papers and presentations on the first day, I found all the sessions very interesting. I don't have a favorite, all the three sessions I attended were very good.

Most of the presentations were given by graduate students. The quality of most of the presentations were very good. It is obvious a lot effort went into the rehearsals of those presentation. Almost all presenters had written (and memorized) extensive speaker notes and while the presentation view was displayed on the curtain, they had the presenter notes open on their laptops. Some of the presenters just read from their presentation notes. Those presentations were not very engaging. But at least the slides were very well organized, and the messages were distilled down to important points and were easy to follow.

Each presentation was about 20 minutes including a 2-3 minutes question answering slot at the end. (I think the SOSP conferences I attended before had 5 minutes reserved Q&A slot, but for this one the Q&A was not as rigidly reserved and enforced.)

Most of the presenters were able to cover around 40 slides in 20 minutes. This is an astonishingly large number. Normally the rule of thumb is to have 10 slides to present in 20 minutes. But being well prepared for a smooth flowing presentation, the presenters were somehow able to pull this off. I guess this takes its toll on the listeners though. I felt overwhelmed and exhausted after three sessions being bombarded by too many ideas, concepts, and acronyms.

I had written two posts about how to present in case you are looking for advice in that department.
http://muratbuffalo.blogspot.com/2017/08/on-presenting-well.html
http://muratbuffalo.blogspot.com/2015/02/how-to-present-your-work.html

No comments:

Two-phase commit and beyond

In this post, we model and explore the two-phase commit protocol using TLA+. The two-phase commit protocol is practical and is used in man...