When someone says: "There's this great dynamically-typed programming language with a useful and expressive type system, powerful macros and multiple-dispatch that solves the 'two language problem' (meaning that you can quickly and comfortably write a prototype in the language but then also write the fast production version in the same language), that has roots in academia but with uptake in industry, a great interactive coding experience in the REPL (including the ability to show you the assembly code for any function!) and [an] excellent compiler[s]" you don't know if they're talking about #CommonLisp or #JuliaLang until they choose either the plural or singular for the word "compiler"! 😛
New fully adaptive Radau IIA method, achieves state-of-the-art performance for high accuracy on highly stiff ODEs. It has a fully automated order construction with adaptive order, and thus if you use higher precision numbers it can automatically construct 17th, 21st, etc. order versions of the method on the fly. Outperforms the classic Hairer Fortran implementation of radau by about 2x across the board!
@bthall (Brandon) and I discussed it a little bit at the time.
I haven't heard anything else about sexual harassment or assaults, so I presume that people's behavior changed.
So a year or so ago, Brandon offered to pay for a year of DataCamp in order to help me prepare for a better job in the future. (Yes, I know I'm old, but I have no pension or other benefits from my previous jobs. I expect to have to work for the rest of my life.) I looked it over and they cover many different things that have been on my "learn this" list for years.
Still working on finishing the track on #DataCamp. But I wanted to add a little more to this.
It took me most of a year to discover this, but I struggled mightily with data analysis functions in #Python + #Numpy + #Pandas, in #R-lang, and in #Julia-lang. #SQL was much easier to comprehend. But I've recently had a few courses where they were covering pure Python, without the data analysis packages, and that is totally different.
Even though I've barely touched Python in the past 20 years or so, it feel familiar and almost everything we do feels "natural". With the data analysis / data science content, it feels like there are dozens of nearly identically-named functions and methods, each with its own special syntax and list of arguments to pass to it.
are easily mixed up and I always (no, seriously always) pick the wrong one first.
I guess that's not a DataCamp issue, but more of a problem with the tools being covered.
But DataCamp's methods don't help with this much. Each one-hour chapter of each four-hour course is supposed to be a sequence of bite-sized tools that one learns to use and then remembers it when it comes up again later. Unfortunately, it quickly turns into a big ball of mud.
@ikonoklast ich will mich an #julialang versuchen. Hab schon ein bisschen Erfahrung und es fühlt sich sehr angenehm an. Vor allem wenn man in Python auf Dinge wie bumpy angewiesen wäre. Julia handelt das syntaktisch deutlich besser, da es in der Sprache eingebaut ist. Es setzt auch deutlich weniger arkanes wissen voraus, dafür allerdings mehr wissen generell. Und die Dokumentation ist vollständig, allerdings sehr unübersichtlich.
This is sort of a tangent to the Google + Coursera https://www.coursera.org/ CyberSecurity and the IBM + Coursera Backend Development programs, but it just happens that both of them are currently covering #Bash and Python right now and the CyberSecurity program is also covering SQL.
My university dropped the campus wide MATLAB license around August last year. It is amusing to see the effect on my GitHub contribution chart. But then I picked up #julialang and now there is more than a recovery :)
I've created a #blog post comparing the performance of standard library 'sort' calls in different languages. Chapel's sort uses composable parallelism to be 10x faster than other popular languages for a test sort on my PC.