A new take on programming trends: You know what's not a bunch of JS? Devs learning Python and Java ahead of JavaScript A report from an online learning platform presents a different take on programming language and devops trends, showing Python and Java ahead of the usual survey winner JavaScript. O'Reilly veep of content strategy Mike Loukides aimed to discover "real trends" which "unfold on much longer time scales" rather than the current …
Unsurprising. Web dev is an exclusionary, pseudo-elitist cavalcade of tools-of-the-month and if you aren't up to the bleeding edge you're immediately out of date.
And the money is pants.
Almost every business on the planet runs on some combination of the JVM, Python and a sprinkling of C#, with some lower-level bits and pieces in the mix for those particularly keen on speed. Everything from your API frontends to your backend risk models - barely a jot of JS in sight once you're past the browser.
Almost every business on the planet
For JVM and Python on the BACK end? You forgot to consider PHP, particularly within a Linux hosting environment.
JVM not so much either... not unless you write Java desktop applications or Android applications.
Python on the backend, seen that - with DJango. i hope I never see that AGAIN.
At least with Java, future devs will be able to do client applications and not just web pages. It's like a stepping stone for a native (read: proper) language like C or C++. Heh.
Python, on the other hand, seems to be way too encumbered with its "scriptiness". For a beginner I'm sure it seems cool, just like BASIC did back in the day. However, for writing maintainable and reliable code that's not overly dependent on 3rd party library HELL, or [worst yet] NOT having some incompatible change made to Python itself, And 'pip' is just a stopgap that hides the weakness, especially when downloading 'the latest' breaks something. And so I do not believe it is quite ready for "prime time".
Still I think Python is great for LEARNING and quicky scripts and prototypes and wrappers for things like GTK and WebKit. But I wouldn't write a commercial application with it
It's also good for demonstrating an algorithm or a process to people who are novice programmers. chances are that if your example is in Python, they will be able to run it and learn from it.
Not surprised it's top of the rankings for a school that is teaching programming.
IME Python is exceedling good for glue scripts and programs that perform 1 or 2 fairly simple tasks that can be done reasonably quickly in a startup/shutdown scenario, , eg unix backend system tasks, DB input sanity checking, archive searching etc. Basically the areas Perl used to be used in before everyone got fed up with the hideous syntax and Perl 6 nonsense.
What Python is NOT good at is large scale backend applications that run 24/7 and have high data throughput, potentially via hundreds or thousands of simultations network connections requiring multiple threads or processes. That sort of thing is best left to C++, C# or Java.
Horses for courses.
> You forgot to consider PHP, particularly within a Linux hosting environment.
>JVM not so much either... not unless you write Java desktop applications or Android applications.
FYI it's 2021 not 1995. Nobody serious uses PHP, JVM is the dominant server application platform and Python is used extensively across a range of backend scenarios with a particular popularity in the areas
of data & analytics (pandas, scipy, pyspark etc), scheduling & orchestration (airflow et al) and configuration management (ansible).
I would compare the proliferation of tools for web development to the Cambrian Explosion - all these weird and wonderful lifeforms with a thousand tentacles or 100 eyeballs or 4 arses that briefly flourish and then die out. Eventually the winner will be some seemingly insignificant little creature that doesn't look fancy at all at first glance but gets the fundamentals right and forms the basis of all that follows.
In the case of web development I suspect that will be WebAssembly but we'll see.
"Web dev is an exclusionary, pseudo-elitist cavalcade of tools-of-the-month and if you aren't up to the bleeding edge you're immediately out of date"
This seems to be what is stagnating it. No real innovations are coming just because nothing can set up serious roots and grow before it dies and gets replaced by something vaguely similar.
Oddly enough the real innovations seem to come from the browser manufacturers (i.e standards like WASM, WebGL, etc) but unfortunately this is starting to fault too with Firefox having a pretty hard time and Google more focusing on extracting advertising control.
But at the same time I don't think the web world particularly needs to get more interesting, it just needs to start cleaning itself up and improving performance and robustness. Perhaps this plateau in browser innovation will be good for it.
They think people should be able to do that with HTML. Because there's not that much risk with HTML. Sure, you could copy a link to something malicious, but it wouldn't serve any purpose and you would change it. You can't do that with the sites which construct themselves from JS, because A) you don't understand what most of that is doing and B) almost certainly the new page wouldn't work, especially if the JS contacts some other system. Their comments are limited to HtML.
Loukides did not specify HTML in his "view source" comment, and I don't see any reason to interpret it that way. Back In The Day, it was quite common for people - professional developers and hobbyists alike - to copy snippets of Javascript and CSS, as well as of HTML, from pages they found.
I personally know a handful of people with no programming background who did that in the late 1990s and early 2000s.
It's a classic hackish approach to experimenting with and learning about the system. I found Loukides's comment completely appropriate.
"Microsoft has "become a leader in open source," he added, thanks to cloud and GitHub"
MICROSOFT has become a leader in open source thanks to... GitHub. Think about this ever spreading opinion. Please, do go on about the "New Microsoft".
Or maybe you're thinking Microsoft opened up Visual Studio, Windows, OneDrive, Office... nope, they still haven't open sourced SHIT yet they are spreading the opinion that they are a "leader" like wild fire when all they have done is BUY the code YOU gave them freely.
Blood pressure reaching critical.... gotta go...
25 years ago you couldn’t write a document in a browser, have it auto save without human interaction, and have multiple collaborators.
That’s just a very basic example. What you can do in a browser now is nothing short of incredible progress. Progress which people who don’t understand how it works behind the scenes get annoyed about. If you’re the sort of person who doesn’t understand why you can’t just copy/paste HTML to get stuff like that to work I can’t help you. The nature of some of these apps has also brought people together to do great things which *is* the very essence which the openness the internet was intended for.
To be fair you're not describing a browser, but a server. Auto saving and multiple collaborations don't happen without a server (which doesn't care about your browser). Also by 1998 at the latest there was IE "plugins" (ActiveX hacks) that allowed you to copy and paste, but you had to save local and usually (always?) to \r\n text with no actual formatting tags (<p>, <br>, <h1>, etc...).
The article was drumming at the fact that today's webpages are just blobs of .js. There's _some_ 3D applications I have to give props to, like fancy wave generators or firework style animations, but 99.9% of webpages are NOT that. Web development is all fucked up to be honest. Somebody, somewhere is hesitant to bring up the fact that modern .js compliance is amazing, probably because they want their blob.js to remain relevant for VERY high commercial interests. None is guilty of that more so than jquery. Using jquery today is a clear sign of ignorance about modern standards and if you're a security buff (I am not), you might also wonder what other standards a jquery user might be ignorant of.
And then there's the re-re-re-re-"invent" crowd... I won't go there. And then there's Java Electron... nope.
Give it time though, it's logically inevitable that eventually someone will port GTK or something to WASM. The Linux kernel and a larger subset of GNU has already been ported, so why not the desktop GUI too I guess (I'm holding my breath for some back to roots WebKit KDE mutant).
I’m actually describing both a browser and a server, amongst other things. Yes you need a server side component to save a document, but the likes of Ajax/async requests (which you need to send the request to save) and support for UI controls built with CSS3 and SVG just weren’t available in browsers that long ago. The slickness of applications now is reliant on everything from the client, server and networks in between. Fault tolerance and redundancy (network / server side) have moved on considerably. In my view it is progress.
@andy I agree that it's remarkable what can be done in the browser today - for proper applications, eg google docs, it's extraordinary. But:
> Loukides is horrified by the way web development has become all about tools which generate screeds of incomprehensible JavaScript.
I have a lot of sympathy with this too. The bar to entry for JS is very low, which means lots of bad design, lots of flailing about on StackOverflow trying to find answers to half understood questions, and lots of shitty frameworks piling obtuse structure and metaphors on top of what is actually quite a decent language. I suspect that's where the "screeds of incomprehensible javascript" is coming from. When I rule the world, one of my first acts will be burning every website that uses JQuery to ashes.
"The slickness of applications now is reliant on everything from the client..."
I think that no matter the interface, it always happens at the server. Using a web browser for example: You click category, you click sort then you click fetch... 3 clicks. What does the server code look like... just 3 obckicks? To be fair, this isn't web browser specific, it's all clients.
Not quite 25 years but not soon after Netscape Navigator had plugins and LiveConnect. And Microsoft had their own Java / ActiveX embedding in Internet Explorer. And stuff like DOM manipulation and Javascript was there. So yes you could embed applications although it was not pretty. I even remember running a beta product office suite written in Java that embedded in a browser, a bit like Google Docs. I think it was Borland who wrote it though I may be mistaken. I also remember a pretty funky VRML renderer and some other cool things all running in the browser.
And looking past how clunky and proprietary these things were, the purpose was pretty much the same as it is now - interactive web content. Fortunately now we have standards but we still have this pig called JavaScript in the middle which explains the proliferation of layers of libraries that are trying to minimize the amount it needs to be used. I suspect as time progresses we'll see more and more complex websites almost revert to the plugin style model except instead of a plugin it will be a webassembly module hooked up to a canvas element.
25 years ago you couldn’t write a document in a browser, have it auto save without human interaction, and have multiple collaborators.
Well, you could have, actually. HotJava was released 25 years ago, and it would have been entirely possible to implement this set of features in an applet.
For that matter, LiveScript was released 26 years ago, and while there was no XHR or WebSockets for background interaction with the server, it would have been possible to, for example, count keystrokes and do a foreground POST when a threshold was reached. The UX might have been a bit clunky - but then that's still true for most of the (horrible) web SPAs I deal with.
Of course, a better question is whether turning browsers into the dumb terminals of the 21st century is a good thing.
I think this is to do with the fact that Wordpress alone makes up a significant chunk of what's on the web. With it being built in PHP it's often hard to separate people from who do actual PHP development, from people who use an application without doing any development at all, that happens to be written in PHP.
25 years ago we had Perl/Asp Classic guestbooks, server side rendered web dictionaries and geocities. Yes, it was nice back then but bemoaning how complex the web is a pointless cliche. Yes, it was lovely back in the 80s when BBC2 ran evening telly programs that everyone with a BBC B could follow but the world has moved on.
Also, these stats are fun for El Reg to publish and for us to guffaw at but they're otherwise useless. The truth is that we don't really know what much of the web runs on.O'Reilly are pretending that their stats are worth more than they are for marketing reasons.
I wonder which kind of domain is the one with the most program lines created per year ?
Sure, the web is the most visible (tautology?), but how about all those embedded programs we do not even see , or do not even know they exist ?
(I read one day there are 4K of code within an electric shaver)
Note this is a true question, I do not have any clue.
But what I know, is that JS did nothing for me when writing firmware for various devices. Neither python, nor java, BTW.
I've "overclocked" 2 battery based electric shavers by upping the voltage, however, some of the newer ones sadly do have micro logic o somer IoT B.S. This makes them useless to me.
Some (most?) of the older battery based beard trimmers can be taken from 9v to 24v (maybe higher, I don't have higher wall warts than 24v), but then you have to bypass the battery thus it's no longer cordless. Why do this you may ask? That's because a lot of the battery powered beard trimmers aren't up to par (by default) with barber shears and barber shears don't typically cut as low as beard trimmers. Typically that extra voltage would be bad with prolonged motor usage, but that extra 13v changes the game for beard trimmers which are only on for 2 minutes top!
If it takes two weeks to learn Javascript but six months to learn Java, then the stats will make it look like Java is more popular to learn.
If it takes a week to write an application in Java but only a day to do the same in Javascript, then the stats will show most development time is spent in Java.
The best metric for evaluating a programming language is some combination of learning curve + maintainability. (Perl is notoriously unmaintainable.)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
