My friends
@lisabdunlap
and
@conor_power23
run a podcast called “Thinking About Thinking About Computers,” chatting with CS PhD students about their research journey & life outside of work. Great resource (especially if you’re considering a PhD program):
Dan Suciu is teaching database theory at Berkeley this Fall and you can find all the lecture videos, slides, and homework problems on this site if you want to follow along at home. Only recordings of database theory lectures I’m aware of on the internet!
Our VLDB vision paper dropped today. We describe how the gap between convergent writes and convergent reads in CRDTs is a "foot gun" and suggest monotone queries as a lens. We also propose many ways database researchers can help! If you want to chat about it just DM me!
Its PhD application season. Here are a few very cool new professors from my lab who you should apply to work with!
Sam Kumar, UCLA: Systems & Security
Mae Milano, Princeton: PL & Systems
Stephanie Wang, UW: Systems
Dixin Tang, UT Austin: Data
Daniel Kang, UIUC: Data & ML
1/2
Programming languages and database research seem to be colliding lately. E.g. CRDTs, Datalog, Egglog, monotonicity, algebraic data models. Are these communities exchanging ideas more and more or am I just in a small bubble of people publishing in both 😵💫
Very pleased to see DBSP win the VLDB best paper award. Algebraic models in databases are red hot right now. Groups, rings, lattices…what will they think of next!
There is virtually 0 Wikipedia content on the major results in database theory. Very little consumable content on these topics on the internet in general.
@justinjaffray
’s blog is one of the only places I know of.
When I started grad school I was lead to believe that whenever someone in the group submits a paper we go out for ice cream to celebrate. I’ve now submitted 10 times and gotten ice cream 0/10 times. Starting to think the ice cream may never come…
Great NULL BITMAP today on the algebraic view of DBs. This is my favorite way to think about systems rn and especially databases and CRDTs. Almost all of my research could be described as “what are the implications of viewing systems algebraically”
Does anyone have book recs on the history of research? I really liked The Idea Factory about Bell Labs. I also read the Xerox PARC one, Dealers of Lightning, but found it much less fun.
So many cool db papers were lost to the ages because they named XML as their application when XML was hot, but then XML became deeply uncool. The ideas aren’t XML-specific!
Watching the College World Series, the players are screaming "lets gooooo" whenever they accomplish an incremental step like a strikeout. This seems fun and I feel like it is really missing from the research experience. Stay tuned for how I work this into my research routine.
Great talks and views at the Sky Lab Tahoe Retreat!
@shishirpatil_
and
@ShadajL
killing it with talks on Gorilla and Hydroflow. Gorilla even supports hydro APIs now!
Had a blast at
@VLDBconf
. Sea otters at the banquet, talking coordination with
@PatHelland
, and a group bike ride through Stanley Park :) Hopefully will see everyone again at CIDR in January!
A funny effect of the very good submit-any-time VLDB deadline model is that when I submit to a conference like OSDI my lab is full of other students working late and you get this nice camaraderie. I'm submitting to VLDB today and the rest of the lab is just having fun without me
These slides are great
@MadhavJivrajani
. Really glad you enjoyed the paper.
@ShadajL
and I are planning to put out some blog posts soon going into more detail on the topics of CALM and CRDTs!
The slides for this talk can be found here:
Thanks so much everyone who attended, and for a great discussion post the presentation, I personally learnt a lot from y'all! ❤️
If you’re a Berkeley PhD admit with questions about Berkeley or the PhD feel free to hit me up. If I answer enough questions I get a free sweatshirt so really you’re doing me a favor.
Found this nice quote by Danny Kahneman that I think really rings true for CS research as well. “I’ve always felt ideas were a dime a dozen. If you had one that didn’t work out, you should not fight too hard to save it, just go find another.”
"Students shouldn't go out into life without the ability to communicate.
Your success in life will be determined largely by...
-your ability to speak,
-your ability to write, &
-the quality of your ideas,
**in that order**."
— Late MIT Prof. Patrick Winston
So Michelin Tires is blogging about Datalog now. 1. That’s dope. 2. I hope people don’t think research in Datalog is too Basic now. The worst possible fate of a research project.
What is the argument against monthly submission deadlines like VLDB? From the student side, these one-deadline-per-year conferences are a much worse experience.
Redecorating the lab for winter with some reimagined database logos. Hydro with a snowflake and Postgres with earmuffs. Lmk if you have suggestions for other winter twists on logos!
I feel like writing textbooks has become much less common for profs in computer science in the last 20 years. Maybe because there are a lot of other well-paying side hustles now like consulting and startups? Thoughts?
Grateful every day for the VLDB monthly submission model. It makes for much more productive research because you can actually serialize deadlines for concurrent projects and resubmissions. Every time I submit to a non-db conference there is some timing collision.
@_TylerHillery
Database engines are pretty darn good. It’s a recurring trend that folks build custom solutions for emerging data problems and then over time those solutions become a layer on top of a relational database.
CIDR NOOOOOOOOOOOOOO deadline extensions are terrrrrrible. Getting a set of co-authors to agree on whether to push for the old deadline or change everyone's plans to go for the new deadline is possibly the hardest part of the PhD.
Love this Terry Tao article on finding appropriately hard tasks for your energy levels. Clearly we need more low-energy research tasks. Research TikToks are a start, but I also want “The Real Postdocs of Palo Alto” or “Keeping Up With the CAP Theorem.”
It’s a common occurrence that I find out random authors, researchers, and podcasters live in Berkeley. TIL Michael Lewis lives in Berkeley! This always surprises me. Like, outside of the campus it’s such a small town. Where are all of these people???
I feel like a ton of great courses were made for an online format and recorded over the past two years, but the youtube algorithm won't show them to me. I have to stumble across them on professors' websites. Has anyone made an aggregator of these by subject?
I feel like academia should make “writing duos” for research more feasible. It seems common to hire a writing duo to write a movie or for a tv writing job together, but we don’t admit research duos into PhD programs or hire professor duos to universities.
There should be a Craigslist-style "missed connections" for the anonymous paper reviewers that seem to like your research but then you never learn who they were.
GSL will host NorCal DB Day at the Microsoft Silicon Valley Campus in Mountain View, on Thursday May 11, 2023. Here is the tentative schedule: . We would like to invite you to this event. This is the registration form: .
@natefoster
@justinjaffray
The Alice Book and Libkin's book are both great for their intended audiences though. There is even a new db theory textbook by libkin and others in the works:
How hard would it be to write papers assuming only what's taught in undergrad CS courses? Add ~3 pages to the page limit for background? It's much easier for readers that know to skip the background than readers who don't to figure out what they're missing and where to find it.
@ben_golub
I was just thinking this the other day. I’ve been preaching balance for a few years now but I realized that it probably isn’t the right advice for someone really early in their career. I was overlooking the years I was grinding that put me in a position to have balance now.
1/2 “creating a class of independent audience-supported scholars and writers and thinkers is likely very healthy for our culture.”
I’m very curious to see how alternative funding models like Patreon, GitHub sponsors, and Substack will change the academic landscape.
@muratdemirbas
@YouTube
We started cross-posting our paper summary tiktoks on YouTube Shorts . Maybe you could make 30 second videos of your blog posts and increase the odds of the algorithm landing on something good!
@justinjaffray
According to wikipedia, logics with least fixed point weren't studied until 1974 (). Made it into database theory by 1979. 10 years late to the relational calculus party.
Very clever how Twitter moved the “promoted” logo to the corner and changed it to “ad” so that I have to expend slightly more effort to ignore ads. They should randomize where they indicate this and what synonym for ads it says on each post. Like a brain game app for ad detection
When I'm iterating on a paper I always print out the drafts and write down the changes on paper before putting them into Overleaf. Is this normal or does this mean I'm old and stuck in my pre-iPad ways?
📢📢 Early bird registration ends today! If you have not yet registered for ACM SoCC'23, do so now!!
Follow this link to register:
@yytian
@ppietzuch
@ozcanfatma44
We have multiple internship positions within our research group! If you are interested in working with us this summer, consider applying for any of the following job postings:
@MadhavJivrajani
I haven’t found any good resources on it really. If you find any, let me know. Would love a good textbook on the topic. I’m hoping to integrate the LTL view of liveness and safety with the CALM work more.
Remember when new CS textbooks weren’t all published as free PDFs? Dark times. I used to have to use this thing called “inter library loan” and wait days.
@justinjaffray
did you know Mark Rosewater recently made a podcast episode entirely about the use of databases in designing Magic The Gathering? Not a single mention of worst-case optimal join.