mratsim

35 ms·

361.
by mratsim 6y ago
Personally I don't want to deal with StackOverflow elitism, "closed duplicate", "you don't know what you want to do, what you want to do is this", etc.
362.
by mratsim 6y ago
I think the best is to go there: - https://nimble.directory/ - https://github.com/Willyboar/awesome-nim - https://github.com/nim-lang/Nim/wiki/Curated-Packages What kind
363.
by mratsim 6y ago
Even as a speaker, I'm surprised by how wide the use cases are being found for Nim: - DSL for real-time audio processing - Embedded programming with light switch - Creating GBA games! - Compiling to Android without Android Studio - Gam
364.
by mratsim 6y ago
In many cases people quit managers
365.
by mratsim 6y ago
Dup: https://news.ycombinator.com/item?id=23573386
366.
by mratsim 6y ago
Reuters read HN?
367.
by mratsim 6y ago
I don't use Go
368.
by mratsim 6y ago
I'd rather have an applied paper have tests, comparisons and source code than lots of theory and being hard to reproduce because "implementation details" don't appear in the paper. Thanks the authors for putting the code
369.
by mratsim 6y ago
For me it's an interface-typed Python. I'm not sure people are even using types, lack of generics and what not, ....
370.
by mratsim 6y ago
Technically robotics isn't mainstream, Javascript/PHP/Java/C# are, at least in terms of developers. Now in terms of code deployed, C on the billions of embedded devices may be able to compete with servers?
371.
by mratsim 6y ago
The logo is so similar to Haskell's ...
372.
by mratsim 6y ago
Very nice website, it's very well done. Regarding performance, I don't see what makes this high-performance. I looked into the code for filters and it's a very naive translation of the algorithms. Naive not in the pejorative
373.
by mratsim 6y ago
Actually on 32-bit it uses a 2^25.5 radix, some are 2^26-bit limbs and others are 2^25.
374.
by mratsim 6y ago
The French schedule is only half the population, the one at 35 hours/week, but many have more hours: - Lunch is often 2 hours for "white collar" workers - Plenty of people start at around 9~9:30AM and finish work at 7:30PM~8P
375.
by mratsim 6y ago
I'm not touching COBOL anymore for $200K
376.
by mratsim 6y ago
It depends, if a potential hire suggests using Python instead of R or Java to process data, that's a reasonable proposal.
377.
by mratsim 6y ago
> COBOL is 65% of active code used today; and runs 85% of all business transactions. Citation needed
378.
by mratsim 6y ago
I'm not sure if I like the conciseness or if I'm leaning more towards "that's all?"
379.
YUV and luminance considered harmful (2008) [pdf](poynton.ca)
2 points by mratsim 6y ago|0 comments
380.
by mratsim 6y ago
Assuming you want a very lightweight HTML/CSS UI engine, I think you can get 80% of the way by using webview[1]. It's very tiny (1200 lines single-header file). I actually wouldn't be surprised if it used webview internally.
381.
by mratsim 6y ago
It's overengineering for sure (or architecture astronauting).
382.
by mratsim 6y ago
It's probably easier to run a JS script to mine crypto than a virtual machine to run a video ad though.
383.
by mratsim 6y ago
You can use them without the HPC expert, Halide for example has a good autotuner and has been used by Google and Adobe to create image filters for mobile devices.
384.
by mratsim 6y ago
Benefits of Open Closed Principle - Extensibility: Not true, unless you solve the expression problem either you pay while adding new types, or you pay while adding new functions - Maintainability: Not true, dressing code in layers of indire
385.
by mratsim 6y ago
That's why you have people working on Halide, Taichi, DaCe, Tiramisu. - https://halide-lang.org/ - http://taichi.graphics/ - http://spcl.inf.ethz.ch/Research/DAPP/ - http:&#
386.
by mratsim 6y ago
The ceremony to start using Java is significantly worse than Python or Clojure. Science is about functions, not objects, a language that requires all functions to be associated with an object is a non-starter.
387.
by mratsim 6y ago
What does Clojure bring? Here is my data science stack https://github.com/mratsim/arch-Data-Science Let's go over it: - Is there a mature tensor/multidimensional array library in Clojure? Is there a Dataframe
388.
by mratsim 6y ago
AI, NLP and ML are bottleneck first by memory accesses and second by CPU. Whatever language you end up using, the low-level implementation must allow you to precisely control memory layout. This is what is being done with C++ + Python but p
389.
by mratsim 6y ago
Seems to me like if you want to look cool nowadays you would use Rust not Python. Different communities have different needs. Data scientists, or physicists, or biologists don't care about memory safety, the programming language is a t
390.
by mratsim 6y ago
> The only one that can strip is the end user compiling a final binary, and the compiler often cannot do this for you because that requires whole program optimization and full LTO, which is super super slow. Whole program dead code elimi

More ›