wdroz

12 ms·

61.
by wdroz 4y ago
> It's a bit ridiculous that setup.py allows you to run custom code on install. In Rust it's "worst", you can run custom code at compile time. But IMO the issue is that we are optimizing everything for convenience at
62.
by wdroz 4y ago
Pattern matching in Python is not perfect, but I still thinks this is a cool feature. The ergonomic doesn't match Rust pattern matching as explained in this recent discussion[0] but it's better than nothing IMO. [0] -- https:&#
63.
by wdroz 4y ago
It's nice to see how much faster is wasm. This kind of problem would also be nice for parallelization. What's the state of multi-threading wasm? any advantage over javascript web workers?
64.
by wdroz 4y ago
I just finished this book. I liked it and learnt a lot, but in some chapters I felt way too far from my comfort zone. The unsafe chapter and some part under the hood of async like Pin/Unpin was difficult for me.
65.
by wdroz 4y ago
I did some Unity3d back in 2013, I don't remember using ECS (intentionally?). After a quick search, it's seem that the MonoBehaviour are the old way and ECS the new ways? That's nice.
66.
by wdroz 4y ago
Even the Data-oriented design[0] is eyes opening (which is part of the ECS). The classic example is[1]: Instead of having a list of structs, where each struct get some types (int, float, ...), you use one struct, where the fields are list o
67.
by wdroz 4y ago
My production docker image with flutter web (frontend) and Actix-web (backend) is 13 MB.
68.
by wdroz 4y ago
I'm doing mosly machine learning/backend. For a PoC I needed a frontend, I choose Flutter web. The tooling works great and is unlikely to change everything week like in the JS world. Dart is easy to learn and hasn't as many f
69.
by wdroz 4y ago
As an European, that baffle me the casual discrimination that are rooted in your US mindsets. The whole "You can do/say X if you are Y, but if you aren't Y, you are a racist/bigot/*phobe" isn't healthy to
70.
by wdroz 4y ago
I think that neither "outlier toolchain" nor "some percentage increase" are fair. This benchmark [0] show significant speedup while lowering the memory needs. You still need to reach dask/spark for really big data w
71.
by wdroz 4y ago
I switched from pandas to polars[0] . I'm happy with the performance gain. You still need to use Spark/dask on "big" data, but polars give you more room before reaching to these solutions. [0] -- https://www.
72.
by wdroz 4y ago
Thanks for your detailed answer. That's sound really interesting. Maybe you can try your luck for "This week in Rust"[0]. [0] -- https://github.com/rust-lang/this-week-in-rust
73.
by wdroz 4y ago
So the GUI is in Qt and the internal engine in Rust. Do you call Rust code from cpp? It's less common in that direction, how was your experience doing it? any obvious traps?
74.
by wdroz 4y ago
In python, I just go with "config.py".
75.
by wdroz 4y ago
Maybe they shouldn't have buy Macs in the first place.
76.
by wdroz 4y ago
If you are unhappy with pandas, give a try to polars[0] it's so fast! [0] -- https://www.pola.rs/
77.
by wdroz 4y ago
> If you create something, make it freely accessible and easy to use, then I think you are partly responsible for its misuse. I don't agree, but people will never reach consensus on this moral topic. So please don't call the ot
78.
by wdroz 4y ago
I wanted to add django but their issues aren't on github as they are using their own system. IMO an external issue tracker lead to a higher barrier to create new issues, so this isn't not necessary fair. Also fastapi allows people
79.
by wdroz 4y ago
It's interesting the see the number of open issues of popular python projects: Pytorch, 8400 Pandas, 3400 Tensorflow, 2100 jupyter notebook, 2000 scikit-learn, 1500 pytest, 720 requests, 170 The outlier seem to be Flask.
80.
by wdroz 4y ago
miniconda is nice for that too, especially if you want non-python things like cuda.
81.
by wdroz 4y ago
The opposite would lead to `(0 * Infinity) === ("foo" / 3)` to be true. I don't know which is better...
82.
by wdroz 4y ago
I hate the nanny point of view of OpenAI. IMO trashing Meta because theirs models may be misused isn't fair. I think that hackers should advocate to have the freedom to toy/work with these models.
83.
by wdroz 4y ago
Fastapi use gitmoji [0], this is well-documented and structured. [0] -- https://gitmoji.dev/
84.
by wdroz 4y ago
But you can read json from polars (or from pandas first then polars) then you can do the map/reduce/filter like in the benchmark.
85.
by wdroz 4y ago
Why they didn't compare with Polars [0]? According to the h2oai benchmark [1], Polars is fast. [0] -- https://www.pola.rs/ [1] -- https://h2oai.github.io/db-benchmark/
86.
by wdroz 5y ago
You are right, I do that too :) It's a bit passive aggressive but more acceptable in professional settings.
87.
by wdroz 5y ago
> I also see this all the time in Email correspondence, you can immediately tell if a person is a native German speaker because they don't capitalize the first letter in an English email after "Dear X, we are writing <blah&g
88.
by wdroz 5y ago
Since wasm supports threads, I wonder if you can speed up these operations further more by using multiple threads.
89.
by wdroz 5y ago
Cobra Kai was the only youtube original I watched. Time to time I checked the "Originals" section, but it's seem all contents are/were about US-centered social justice stuff...
90.
by wdroz 5y ago
Apparently yes, sound like the wage gap - differences between individuals of a population are larger than between populations in the first place - differences between populations are largely not caused by *sexism*

More ›