[HN Gopher] Why Julia Is Turning Heads in 2021
___________________________________________________________________
Why Julia Is Turning Heads in 2021
Author : tmfi
Score : 29 points
Date : 2021-04-04 17:51 UTC (5 hours ago)
(HTM) web link (www.nextplatform.com)
(TXT) w3m dump (www.nextplatform.com)
| 0-_-0 wrote:
| Only on HN is it obvious that "Julia" in the title refers to a
| programming language :)
| elzbardico wrote:
| It is not. There's only this constant spamming done by julia
| advocates on HN. It is getting annoying TBF
| internetslave wrote:
| For years I've seen these hype Julia articles. Yet, out of all
| the open source, all the companies I've worked for, I've never
| come across any valuable system running Julia. I've never needed
| open source and found that only Julia serves the need. I'm kind
| of tired of programming click bait that ends up on HN. Much like
| the "front end programming has failed" article that is popular
| right now. It's all for clicks. The reality is mainstream
| programming paradigms are effective and work.
| jakobnissen wrote:
| It's really hard for a language to grow mature enough for
| businesses to use it for valuable systems. You don't just need
| the programming language, it's also expected you have good
| editor support, some type checking tools, performance
| measurement tools and a whole bunch of battle-tested packages
| for common problems in programming. It's also risky as a
| business to begin building stuff in Julia when it's still a
| relatively niche language. That's not management being
| reactionary, that's just being pragmatic.
|
| Almost by definition, new languages just _can 't_ compete with
| established languages on maturity. They have to compete on
| other metrics - better language design, a different runtime,
| new capabilities, etc. and therefore will attract a different
| crowd - initially
|
| Over time, the bleeding edge first-movers will create basic
| necessities, which then attracts the programming language
| afficionados, who build interesting new software and nice-to-
| have, which attracts ordinary programmers, who then grow the
| community enough to give it stability and battle-tested
| packages, which then lets the language become "boring, old and
| stable". But this whole process takes ages. Look at e.g. Rust.
| Julia is somewhere between the "afficionado" and "ordinary
| programmer" stage. Of course you shouldn't use it for mission
| critical software where maximal stability is key. That doesn't
| mean the Julia hype isn't warrented.
| steveklabnik wrote:
| "Crossing the Chasm" is a classic business book that explores
| this idea. It's about technology adoption more generally, but
| I believe it applies to programming languages too.
| derstander wrote:
| Julia 1.0 was released in late 2018. I think it would be
| irresponsible to design valuable or mission critical systems
| with a language still undergoing that much change. I expect
| we'll start to see the sorts of valuable systems you refer to
| start to surface in the near future if Julia is getting
| traction.
|
| Full disclosure: I don't currently use Julia but I'm interested
| in spending some time with it this year. I primarily use MATLAB
| in my day-to-day life. I like Python as a general purpose
| language but find the syntax of Python+NumPy a little clunky
| for what I do.
|
| I don't really feel that Julia falls outside of mainstream
| programming paradigms anymore than something like MATLAB.
| satya71 wrote:
| Haha, I switched away from MATLAB to numpy many years ago
| because MATLAB was so clunky. To each his own, I guess.
| currymj wrote:
| do you work on high-performance computing, or more generally
| with numerical codes for scientific/engineering simulations?
|
| this seems to be the area where Julia is getting the most
| traction.
|
| it's strong in some other areas too. it's not really being sold
| as a general purpose tool for building business systems,
| though.
|
| (not that it's unfit for that purpose -- just it doesn't have a
| strong advantage. whereas for numerical codes, optimization,
| and a few other things, it is actually better than most
| alternatives.)
| jll29 wrote:
| I tried Julia in 2018 as my new programming language to learn
| that year. The syntax is clean and simple, and I remember it as
| a "Pascal for scientific computing" - as easy to read as Python
| yet faster execution.
|
| I liked the static type checking and compilation, but found it
| frustrating at the time that the sample code from a book I had
| purchased didn't work; overall, the language seemed to be still
| too much in flux to be entrusted with an important project,
| although I have spoken with fellow scientists who use it as
| their main language for experimentation and are happy.
|
| Python is clearly winning because most important libraries are
| C++ under the bonnet - or even Cython (which is used e.g. by
| SpaCy), and its ecosystem with IPython, Jupyter notebooks and
| comprehensive data science libraries (pandas, matplotlib,
| sklearn, numpy, ...).
| threatofrain wrote:
| https://pypl.github.io/PYPL.html?country=
|
| https://redmonk.com/sogrady/files/2021/03/lang.rank_.0121.wm...
|
| https://trends.google.com/trends/explore?date=today%205-y&ge...
| codemac wrote:
| Did this article ever answer the question?
| adamcstephens wrote:
| It's a statement not a question.
___________________________________________________________________
(page generated 2021-04-04 23:02 UTC)