kmicinski

5 ms·

1.
by kmicinski 4mo ago
This just kind of sounds like a random idea that sounds good in your head but not based in reality; the point of a PhD has always been one thing, and one thing alone: train someone who can publish influential papers in top-tier venues. Anyo
2.
by kmicinski 4mo ago
> The US used to be the most desirable place for immigrants and the US higher education system used to be the envy of the world and now for both of these it is not any more. That is a nice sound bite for TV--the reality is that the gover
3.
by kmicinski 5mo ago
This is very cool to hear. Please get in touch with me, I would love to learn more. By the way, I am recruiting participants for an upcoming seminar in which I am soliciting industrial participation: https://kmicinski.com/mi
4.
by kmicinski 5mo ago
I am the author of this paper, and I do not agree with Dr. Smaradgakis' comments. As far as I can tell, the root of his concern is that that paper did not target Souffle Datalog, a specific Datalog language in which his group writes. T
5.
by kmicinski 10mo ago
> In an introduction course there is no need to aim for maximum performance if parsing a 10k lines program takes less than a second. I'll do you one better. The main compiler in my course uses only six characters to parse every sing
6.
by kmicinski 10mo ago
Respectfully, I think what you mean is that there are no books which give you the experience of hacking on LLVM for several years.
7.
by kmicinski 11mo ago
It's true that Ascent does not index ADTs either, but there are some tricks that you can use when you control the container type to get similar performance by, e.g., storing a pre-computed hash. I believe Arash, the main author of Asce
8.
by kmicinski 11mo ago
> - I had worse performance with Soufflé than Ascent for my program for some query-planning reason that I couldn't figure out. I don't really know why; see https://github.com/souffle-lang/souffle/discu
9.
by kmicinski 1y ago
I appreciate your thoughtful criticism of the post, to my eyes everything you are saying is correct.
10.
by kmicinski 1y ago
Seems potentially interesting to explore what would be required to store durable continuations. Feels very related to incrementalization and provenance, as you can see materializing a continuation to disk (whatever storage backend) requirin
11.
by kmicinski 1y ago
I agree with you--that's a topic I will definitely cover in my blog, too. You make a good point: I know some folks who worked at big financial orgs, writing hundreds of thousands of lines of code, and never wrote general-recursive fu
12.
by kmicinski 1y ago
I completely agree with you, but this event was held in October, 2024--Trump's horrendous behavior wasn't an effect on that. There are other reasons folks internationally probably don't want to travel to Texas aside from Trum
13.
by kmicinski 1y ago
For materialization-heavy workloads (program analysis, etc.), we often find that optimized binary join plans (e.g., profile-optimized, hand-optimized, etc.) beat worst-case optimal plans due to the ability to get better scalability (less lo
14.
by kmicinski 1y ago
No offense, but I wouldn't take Datalog 2.0's small attendance as an exemplar of Datalog's decline, even if I agree with that high-level point. Datalog 2.0 is a satellite workshop of LPNMR, a relatively-unknown European confe
15.
by kmicinski 3y ago
Yeah, IncA (compiling to Souffle) and Ascent (I believe Crepe is not parallel, though also a good engine) are two other relevant cites here. Apropos linked data, our group has an MPI-based engine which is built around linked facts (subseque
16.
by kmicinski 3y ago
Right--but CTEs are orders-of-magnitude slower than Datalog, to the point that they are not seriously worth considering for any modern application where Datalog would be used. As you said, Datalog is less expressive than SQL's ability
17.
by kmicinski 3y ago
Yeah, there are a ton of substantively different approaches to modern Datalogs, targeting different applications. To start off: Datalog is distinguished from traditional SQL in its focus on heavily-recursive reachability-based reasoning. Wi
18.
by kmicinski 3y ago
Sounds awesome--feel free to get in touch with us (the authors of this paper) and share your progress. We have a similar single-node Datalog engine in Rust, it would be cool to benchmark your results against parallel Ascent ( https:/&#