What is the current new hotness among hipster web frameworks? I currently use Rails professionally, but being aware that I'm working for hipster business logic processing stuff means that it's going to get replaced with some other nonsense in the near future.
So what is it? AngularJS? Ember?
>using web frameworks
Code it yourself, lazyass. It ain't hard.
>>45317130
meteor
>>45317140
>hey boss, let me create a web framework from scratch! it ain't hard!
>please anon not this again please we're already weeks behind
>>45317130
Did you just put rails and Angular in the same basket? You dont make money programming.
As to your question Node is the hot things right now
>>45317150
I thought everyone makes fun of meteor?
seems like it's still node.js, ember, angular.
>>45317193
Yeah, as a matter of fact I did. They're both web frameworks. Problem?
How does /g/ feel about Flask? I've been using it for a few weeks now and it's really starting to grow on me.
>>45317193
A rails API backend and angular frontend is fairly popular
https://github.com/phoenixframework/phoenix in future
>>45317285
>Python
>>45317231
Yes absolutely,
Rails and Angular are as related as ASP and jQuery
>>45317167
>hey boss let me fit this round peg in a square hole
>>45317167
>>45317339
Do both Rails and Angular depend on the web? Yes? Then both are web frameworks.
>yfw adults are talking and you get told to go home
I also work in a ruby/rails shop
The stack we use is postgres + resque + rails + backbone
But angular seems more popular now
>>45317130
>Rails
Too damn slow.
>>45317430
OP here.
I think web frameworks are bullshit personally, but frankly I don't care enough to go against the grain on this one. All my superiors love them and think that finding new and creative ways to display text on a screen is technology advancing.
>>45317231
>>45317454
>All my superiors love them and think that finding new and creative ways to display text on a screen is technology advancing.
I totally believe that, it just has nothing to do with performance when using a system in production.
>>45317481
Yeah I've sort of resigned myself to frameworks being slow. There's a very real problem among programmers that insist on using every bit of the system resources they're given as hardware improves but have no real intention of improving the actual output of the fucking program itself in the process beyond minor visual enhancements.
Where do all these resources go? Into layers and layers of overengineered, pointless abstractions that serve no fucking purpose other than to make the code 'clean', i.e., pretty
>>45317551
Yeah well you probably would develop for web in C or assembly for speed!
So I'm not going to debate you but using frameworks actually speeds development by quite a bit and since man-hours are worth more than hardware it's a tradeoff that most people and actual money making companies are willing to make.
NEETs can take all the time they want developing for CGI
>>45317167
>someone suggests you don't use a web framework
>you are so convinced that a framework is necessary that you think he's telling you to build your own web framework
lel
>>45317625
Slippery slope argument doesn't work here.
Writing pages in HTML (limited or no js) and python is quite reasonable and results in high performance without slowing development.
>>45317654
>Code it yourself
???
Groovy on Grails, no more fighting gem package manager clashes, everything in the JVM ecosystem 'just works'. seamlessly integrates with existing Hibernate framework apps
Node, Angular and Mongo is what bearded Mac people around me blabber most about... .NET might have gotten a boost with the whole open sourcing and people here have a fetish for it anyways. (I am aware that all of those things are completely different tools for different jobs, I was just generally naming stuff that seems "hot right now").
>>45317625
Depends on the project. If you're writing a program that is some variation of insert content here, generate text based on that, then yeah, abstract away.
It makes sense for webdevs to depend on frameworks. I just aspire to be more than that one day, which means becoming an "evangelist" for some framework or other is stupid IMO.
>>45317686
>python is quite reasonable and results in high performance
lol no. Python frameworks are typically slightly faster than Ruby frameworks but they're still ridiculously slow and they both fail under a moderately concurrent load.
>>45317784
OP here. Pretty much this. I love how employers fork over tons of money in hosting fees just so they can have their juggernaut servers can query a database and dump out a bunch of text.
>>45317731
Yeah I'm no evangelist for any framework but definitively I'm an evangelist for using *some* framework if you are doing anything heavier than displaying some static HTML
Even minimal frameworks such as Sinatra and Flask are a great help for web devs and speed development by a great deal
>>45317850
I mean, I think the issue is how important number crunching power is to the project. If you have to crunch numbers, you have to use a real programming language like C or maybe even ASM, depending on how fucking hard that thing needs to go.
If it's your average insert-content-here website with some ad hoc variation on it, yeah, I agree, Rails or whatever is worthwhile. Unless you plan on having more than five people on your website at once, in which case, you should probably use something simpler and deal with the overhead of actually writing it.
>>45317285
I use flask in production, works like a charm.
Made some tools to automate project setup process and stuff.
>>45317710
It's the first MVC framework I used, it's really nice.
>>45317130
There's a huge motherfucking backlash against AngularJS, for very good reasons.
Don't use it.
>>45318940
What are those reasons?
>>45318940
This. AngularJS is really awful and the current version is going to be completely unsupported in two years and replaced with a completely incompatible Angular 2.0 which is written in yet another new programming language that compiles to JavaScript (AtScript). Just learn Knockout or React with ReactRouter and Reflux.
>>45317784
Flask with uwsgi emperor can handle 10,000 users on at the same time
>>45318940
where? why?
I'm also trying to figure what crap I should say I have experience with.
It's funny how in university we're taught how to do shit the hard way with all the whys and hows but all internships and entry tier code monkey jobs available for no name kids only care that you have experience with this or that framework, tool, development method, etc.
>>45318987
http://larseidnes.com/2014/11/05/angularjs-the-bad-parts/
http://blog.dantup.com/2014/10/have-the-angular-team-lost-their-marbles/
http://www.fse.guru/2-years-with-angular
Use React.js
>>45318983
Also don't use front-end routers.
Don't do single-page apps for the love of god
>>45319020
>Don't do single-page apps for the love of god
>implying we have a choice
>>45319002
yeah tough luck real life is not the same as school
>>45318986
Feel free to benchmark that with wrk. Even on great hardware with 8+ child uwsgi processes Flask and uwsgi will buckle way before that.
You can get that kind of performance with substantial caching but that proves absolutely nothing.
>>45319020
Nah, single page apps can be great as long as they use the history API properly.
>>45318977
> extremely poor performance
> destructive api changes for no decent reasons
> google doesn't dogfood it
> fucked up scoping rules
> too many unjustified novel concepts
> broken modularity
> de-facto *incompatibility* with browsers that can't parse HTML with XML namespaces
> doesn't encourage good practices, *do* encourage bad ones
Use React.js
Are you guys CS majors?
Isn't web dev kind of for people without degrees or CS failures?
Do hipsters use JavaScript frameworks ironically?
>>45319081
>Are you guys CS majors?
>Isn't web dev kind of for people without degrees or CS failures?
OP/CS major here.
Yes. Webdev is for those people. That said, I have to pay my rent while I'm in school, so...
>>45319082
We're doing so because we're crumbling under the weight of the fathomless horror we have created with jQuery.
>>45319015
>This is insane. It should be an uncontroversial statement that one should be able to understand what a program does by reading its source code. This is not possible with the Angular DSL, because as shown above a variable binding may depend on the order in which a user interacts with a web page.
Whoever wrote that needs to learn some Common Lisp. According to Lisp programmers, these are good properties for a language.
>When scope is dependent on program state it is called dynamic scoping. Programming language researchers seem to have figured out pretty early that dynamic scoping was a bad idea, as hardly any language uses it by default.
The entire Lisp community thought the opposite before Scheme came along. They said dynamic scoping matched the Lambda calculus and lexical scope was an efficiency hack from Algol.
>>45319114
>dynamic scoping matched the Lambda calculus
What the fuck kind of lambda calculus have YOU been reading? Because the one I know absolutely uses lexical scoping for everything.
>>45319114
Basically everyone, including modern Lisp communities, agree that dynamic scoping is idiotic and extremely error prone.
>>45319081
I'm a CS major I used to work doing lower level C++ for a banking software company
I changed to web dev because i wanted more freedom, more jobs, and just better work conditions than being in a closed space with three other neckbeards, now i work in many different projects, sometimes from home and have an all around happier life, yes a lot of what I studied for CS goes unused and you might call me code monkey but a happy one
>>45319114
Once I read that JavaScript programmer are judged by how well they understand JS scoping rules.
That was pretty hilarious because I had just realized that JavaScript had totally normal scoping rules but that everybody got confused by the introduction of closures into the mix.
But regardless.
Just the part about two-way biding (and the horrific way Angular does it) gives me the creeps.
>>45319090
>That said, I have to pay my rent while I'm in school, so...
As someone looking to get into backend, do I need to worry about CS graduates taking mah jerbs?
I'm going to a community college for MIS and spending 6-8 hours a day teaching myself web dev.
I'm 30 years old and went to school when I was younger for business, so some of those credits easily transferred over to MIS. It's only an A.A.S. and I really don't have time to start from the bottom with the maths required for CS.
They don't have classes for anything related to backend web development, but I did take web 1 and 2(css, html, and very little javascript, but I learned what was left out on my own) and CS 1 and 2 (java) , so I understand how to program and create websites.
Does my plan to get into backend web development sound plausible? The idea is to just learn backend languages in my free time
>>45319081
When you're fresh out of education with nothing on your resume except where you graduated and your shitty github projects you have two options:
>web dev
>maintaining ancient enterprise systems
There's also nepotism I guess
>>45319181
If you work hard you'll be fine. Actually make things and post some of your code on Github so potential employers can see it. Getting a job in programming is easy if you're even slightly competent.
>>45319196
>and your shitty github projects
If you actually know how to code and make things in normal languages (not Haskell, Lisp or other languages no one cares about) you'll get contacted quickly based on Github contributions alone. I posted a few Python projects on Github and was contacted by tons of companies, including Google, within a few weeks.
>>45319140
>JavaScript had totally normal scoping rules
JavaScriptvar sum = 0;
for (var x = 0; x < 10; x ++) {
for (var x = 0; x < 10; x ++) sum++;
}
console.log(sum);
C/C++#include <stdio.h>
int main() {
typedef int var;
var sum = 0;
for (var x = 0; x < 10; x ++) {
for (var x = 0; x < 10; x ++) sum++;
}
printf("%d\n", sum);
}
The loops are character-for-character exactly the same.
Now, without running either example:
What's the answer?
>>45319268
Who cares about block scope.
Syntactic blocks are a relic of the era where explicit jumps and gotos had to be eliminated without too big a performance hit.
Functions should be used instead.
Arguing about which JavaScript framework is best is like arguing about which animal's feces you would rather eat.
>>45319354
So.
How do you write use interfaces?
>>45319354
We should all use J2EE and abuse runtime annotations and dependency injection to be truly enterprise.
What is wrong with plain old jquery?
Seriously I don't do web dev beyond one page prototypes, enlighten me.
>>45319369
Haskell and ncurses.
>>45319081
Is this a troll? Outside of the 25 people on the planet creating elite C++ software it seems like everyone works in webdev cause that's where the money is. There are so many clueless enterprise java retards who think they're the "real professionals" that I honestly can't tell if you're trolling or not.
>>45319285
Not sure if JIDF (JavaScript Internet Defense Force) or just Reddit/Hacker News parody.
>>45319381
It's very, very hard to write self-contained modules with jQuery. Plus taking a piece of server-side HTML and having the JS act on it after the fact is, like asking for bugs.
The main advantage of frameworks, I'd say, is data binding. DOM generated and updated directly from the data, without explicit DOM manipulation.
>>45319387
look at the Bad Man we have here. He's so Bad. I want to fuck him all night long.
>>45319433
I can now defend JavaScript with the same arguments I use to defend functional programming languages.
I genuinely like this situation.
And I almost always use jQuery forEach instead of for loops. Fuck loops.
>>45319425
Just ignore every Java developer that unironically uses Spring. That eliminates like 90% of the Java community.
It's amusing how so many Java developers proclaim to like static typing but actively avoid using it every chance they get.
>>45319470
>Fuck loops.
>>45319485
>It's amusing how so many Java developers proclaim to like static typing but actively avoid using it every chance they get.
this
OOP in the Java sense is not, by any fucking means, a good example of static typing.
>>45319500
And what is?
>>45319434
Thanks for your answer, that makes sense.
>>45319495
Why should languages have both functions and control flow primitives while functions can perfectly emulate control flow primitives as soon as you have first-class functions?
Block scoping, if blocks, loop blocks are all deprecated versus pattern matching and map/fold/filter.
>>45319512
Off the top of my head, F#, almost any language in the ML family, Rust and Haskell. I'd say Scala and C# are also passable. If C# had proper sum types I'd probably consider it great.
What's a web framework?
>>45319512
HM typed langs.
>>45319518
I really meant that. The original paradigm of javascript is that it's a way to manipulate an existing data structure, but once you go past a certain compexity treshold, it starts falling apart, because the front-end behavior starts being way, way too dependant on what's generated by the back-end. The situation that happens to my buggy code all the time is that, while manipulating the DOM, I end up in the situation where I need to *generate* dom, which was originally the task of the back-end. At that point, the slightest mismatch between the DOM generated by the both generators is a source of bug, besides being an obvious duplication of efforts. It already makes sense to let the JS be completely in control of the final DOM, or at least, where you have heavily interactive interface elements. This is what I'm doing. This is what *facebook* is doing with React.
Now, the other cool thing is that...
Well....
>>45319470
It all makes sense now. JavaScript is just Lisp with more readable syntax even though the syntax isn't always the same meaning as C. But it shouldn't be because it's not C, it's Lisp.
All forms of dynamic typing is shit. At least get shit you can read. That's JavaScript: shit you can read.
>>45317130
code C and stay calm
>>45319603
Something dumb programmers make for dumber programmers to use.
>>45319645
>string manipulation
>C
No thanks
>>45319648
I'll never understand web dev
>>45319609
HTTP/HTML is stateless. Was so before JS, anyways. Shoot HTML on a page, and it won't change before the next request, where you get to rewrite it front scratch.
This sounds like functional reactive programming, which amounts to say that an piece of UI is generated from a given piece of immutable data, and that manipulation of the UI itself is prohibited.
The data becomes the single source of truth.
Changing the data, or rather, replacing it with a completely new piece of immutable data, amounts to notionally re-generating the whole UI element from scratch.
The data thus becomes the source of truth in the application, so it's impossible for the UI to become out of sync with the data, because the UI depends completely on the data.
This is like replicating the original HTML MVC paradigm or never having to explicitly *change* a UI element, only in the browser.
>>45319058
>>45319015
Angular developer here, your posts are pretty BS, here's why:
>complain that angular 1.x has problems, then later complain that angular 2.0 is doing too many new things
>not realizing that all web technologies get deprecated after 1-2 years
The current version has some problems, 2.0 is going to be a massive improvement and simplification using ES6 features.
The one thing I agree with in one of those articles is that you NEED a lead front end developer to manage an Angular project otherwise it's going to explode. This is just the nature of single page apps, there are a lot of moving parts and pieces of state that need to be managed. I personally haven't found a good way around this yet.
React is not a valid alternative, if you add enough parts to it to match the feature set of Angular it starts to look a lot like Angular and have the same problems.
>>45319622
I'm personally opposed to dynamic typing, but at the same time, the LISP community, before the 90's, is responsible for such an incredible amount of extremely relevant research, it's hard not to be a fanboy.
>>45319726
But why are you making single-page apps *in the first place*
I recently read an article about how Angular isn't well suited for large serious projects, but I can't seem to find it now unfortunately.
Currently Go and Node.js seem to be the new big thing. For Node.js you've got frameworks like Express.js and Sails.js (Sails.js being the most reminiscent of Rails).
>>45319622
>JavaScript is just Lisp with more readable syntax
not really, javascript can't into recursion without stack overflow
>>45319622
>JavaScript is just Lisp with more readable syntax
This is completely false.
>>45319745
Because the boss think it's cool and trendy
>>45319521
It sounds like you drank too much JavaScript Kool-Aid. Soon the cult will have you believing Brendan Eich invented the lambda calculus and JavaScript inspired John McCarthy to put the lambda into Lisp.
>>45319745
The same reason anyone makes apps for the web at all, ever?
>>45319747
I was mostly talking about the fact that JS, like LISP, have easy closures, a simple, universal and easily stringifiable data structure, and dynamic typing/interpretation.
There are very few languages which have this apparently simple and minimalistic set of basic features, somewhat surprisingly, and obviously none of them was forced to become widely used the way it happened with JS.
And now you see people saying that the "good parts" of JS are those that look like what LISPers formalized in the 80's, and you see interpreters working the same way LISP and Self interpreters do.
The parallels are easy to draw.
Even though JS has no TCO. How sad.
>>45317140
No one does this in the real world, in any language.
>>45317130
op you are a hipster, because you get money from hispter framework and work for hipsters, in fact I think that is worse than being a hipster
>>45319804
Can you just stop spilling all that jelly all over the place for one second? You're making a mess.
Just do it in C.
>>45319784
All of these describe Perl too.
I doubt you would try to argue that Perl wasn't widely used.
>>45319726
Same poster here, I lied, the way around this is good testing. What there isn't a way around is the need for good testing, but this applies to every large program. It seems that too many frontend developers are not using to writing tests, React is not a solution for this either.
>>45319745
Are you denying that there is a real demand for single page apps? Because I would strongly disagree with you there.
>>45319874
Perl auto-flattened trees into lists.
Case closed.
>>45319521
JavaScript is getting block scope like C has, like ML has, like F# has. When it does, you will begin to use block scope and you will pretend broken scope "hoisting" doesn't exist. All of these arguments will vanish and your Stockholm syndrome will drive you to develop new ones for the many remaining shortcomings of the language. This cycle will repeat over and over again, with you adopting more and more ideas that you once dismissed. I know this reaction from the Lisp community, a near parodic blend of NIH syndrome and condescension.
>>45319754
Actually you are both wrong it scheme
Learn your history children
>>45320166
YOU SUCK
sorry i just felt like saying that
I don't get this client side rendering horseshit in general.
The whole reason for existing was "muh performance". So AngularJS comes along and everyone's like "my answer to our performance prayers".
And then as times goes on, there's blog after blog about "how we dealt with performance issues with Angular", which frankly was never a problem to begin with in traditional rails environments.
>>45319090
Ya really. Franly web dev seems to be where the glamour is. It's like when there's an engineer no one cares about but the salesman who sells smooth talks and sells shit is the company hero.
Well the web dev makes the website that markets the company, he's basically the new salesman.
>>45320393
Angular has performance problems with really large complex pages because it uses dirty checking. Most client side templating solutions (Knockout, React, Ember for example) don't use dirty checking and thus don't have those performance problems.
> which frankly was never a problem to begin with in traditional rails environments.
kek. Rails is slow as shit and with a few thousand users basically everything becomes a bottleneck. A Rails app that does nothing but respond with a plain text "hello world" doesn't even meet Googles or Facebooks strict performance requirement for response times.
>>45320402
You sound like a disgruntled moron. It's far more likely that you're autistic and yet another example of the Dunning-Kruger effect. There's probably a Haskell/Lisp thread up where you can show off your 1337 fibonacci skills.
If you want to be a real hipster, you better start learning scala and use the Play Framework
>>45320482
Even hipsters realize that Play is shit.
>>45317710
This seems in for people used to developing with Spring MVC
>>45320575
Fucking hell. Everyone that supports Spring or Groovy should be fired and never allowed to work in this industry again.
One thing I like about Google is that they respect the JVM but they ignore, and often hate, 99.9% of the garbage created by the Java community. Just mentioning garbage like Spring in a positive light is enough to get you fired.
>>45320633
Take a moment to feel sorry for all those devs stuck working on Spring projects that'd rather have nothing to do with web dev at all
ASP.net vnext. Versatile, Crossplatform, great language and great use of that langauge.
>>45319090
In what planet is web dev for "CS failures"? You must be an unemployeed NEET neckbeard to not know how much money people pay for web development, and the prereqs are 4.0 from a top school in CS.
>>45320633
What's so bad about it?
>>45320633
What's the problem with it?
>>45320651
I'm an unemployed NEET loser and I'm currently learning a MEAN stack - reading that post hurt inside because he's absolutely right about it being for failures.
>>45320653
>>45320670
Spring basically forgoes static typing at every possible opportunity. For some reason, Spring users prefer to be programming in a disgusting mix of XML and runtime scoped annotations that push as many errors as possible to runtime. It makes even the most trivial shit virtually impossible to debug with layer upon layer of terribly inefficient and unnecessary indirection. Even the most trivial Spring app pulls in like three dozen dependencies and probably half of those are used for retarded runtime reflection hacks.
Also, if your Java code is as slow as Python you're doing something horrifically wrong. Just compare Spring's performance against something like Dropwizard, standard Jersey 2 or Spark. It's fucking disgusting. http://www.techempower.com/benchmarks/#section=data-r9&hw=peak&test=query
>>45320773
Thanks
>>45320500
No its actually pretty good.
>>45320816
> reflection based routing DSL
> global application object
> terrible Action and Form API because they want to avoid scary concepts like Functor, Applicative, Validation and Monad
> end up reimplementing all of the above concepts in far worse and far less general ways
> anorm does additional queries to get database metadata (I think this has been fixed but it's just so ridiculous)
Basically everything in the Scala ecosystem outside of Typelevel libraries and Colossus is complete and utter shit.
>>45320633
>Fucking hell. Everyone that supports Spring or Groovy should be fired and never allowed to work in this industry again.
>hipster raging against every major corporation that uses Java enterprise frameworks instead of his flavor-of-the-week framework
>>45321168
Nope. I don't mind Java at all but Spring is objectively awful. Dropwizard, Jersey, Undertow are all great. Spark would be decent too if it got rid of the global routing map.
>>45318986
It's probably the same bullshit as with tornado which only can handle 10k longterm, non-realtime connections.
>>45318986
As soon as you do something more than hello world, the number will drop to 200.
>>45319470
forEach is slow as fuck
>>45319521
>Block scoping, if blocks, loop blocks are all deprecated versus pattern matching and map/fold/filter.
OK kid.
