[ home / board list / faq / random / create / bans / search / manage / irc ] [ ]

/prog/ - Programming

Programming board

Catalog

Name
Email
Subject
Comment *
File
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Embed
(replaces files and can be used instead)
Options
Password (For file and post deletion.)

Allowed file types:jpg, jpeg, gif, png, webm, mp4, pdf
Max filesize is 8 MB.
Max image dimensions are 10000 x 10000.
You may upload 1 per post.


File: 1421771585722.png (1.55 MB, 1366x768, 683:384, 1409640772840.png)

37b299 No.1082

Share your unpopular opinions. I'll start:

I think Python is the best general purpose language if speed doesn't matters.

bd835c No.1086

>>1082
But that is true, specially when we are talking about low level programming like image processing, data compression, and so on. Of course, always programming in Python without any kind of bindings.

My unpopular opinion. Ok, we can agree that most time, Stallman is a faggot, but in most cases, he is right on things he says.

c59a6b No.1089

>>1086
He's a disgusting sociopath with a brilliant mind.

8da8b2 No.1126

>>1082
s/Python/Perl/

c5c073 No.1128

>>1082
Python is also great for beginners, IMO. Easily kicks the shit out of assorted crap like Robot Karel

b57f8d No.1133

Python is shit

680735 No.1135

>>1133
enlighten me

56305d No.1136

HTML is a programming language.

de8ea7 No.1137

>>1135

poor performance
gimped lambdas
shitty oop
code manipulation and generation sucks
guido is a retard

de8ea7 No.1141


ab4f91 No.1161

>>1136
HTML+CSS is Turing complete, so it's a language to me.

>>1137
What should I use then? For programs and scripts that speed doesn't matter

265ac2 No.1169

>>1082
Where the hell is that an unpopular opinion?

>>1136
I can't agree with that, they're markup languages. They formally describe the content of a page, that's it.

537a3b No.1178

>>1082
oop is shit

ab4f91 No.1181

>>1178
Lisp, Scheme, or Haskell programmer?

fcbabb No.1183

>>1137
>>1133

I fully agree with this. Python is like a joke that caught on. It's the most barebones language you can possibly learn. The only thing that saves it is the huge amount of libraries.

1f4103 No.1184

>>1161
It's a mark up language and deals with display rather than computation, but I wouldn't say it's a programming language. As far as I know about HTML & CSS, it can't create interpreters of it's own language like programming languages can.

b8f082 No.1192

Python was my first language. And when I made the transition to C++, I had no idea why everything is so needlessly complicated. So yeah, beside performance, and library support in other languages Python is the best.

c66ede No.1195

File: 1422907398806.jpg (36.11 KB, 300x440, 15:22, 1421090819848.jpg)

>Not choosing the best scripting language

537a3b No.1196

>>1181
haskell

ad605c No.1258

>>1195
This guy gets it

ad5ac3 No.1298

>>1178
These are supposed to be unpopular opinions.

722d37 No.1311

>>1183
>It's the most barebones language you can possibly learn
That's probably why it's adopted so widely. The people who don't want to learn can have their cake and eat it to some extent.
>>1195
Ruby isn't much better than python. To be honest it's hard to name a good scripting language. If the syntax were a little less obscure I'd say perl.

My unpopular opinion? I don't know, probably that bourne shell is a comfy language.

97b661 No.1313

>>1311
>people who don't want to learn
There's nothing wrong with learning an easy language before graduating to a more complex language. I started in Python, moved to Javascript, and now I'm learning C++.

So I guess my unpopular opinion is that Python is a great language that can be used in real world applications. IIRC, Dropbox uses Python, as well as Industrial Light and Magic. I recently saw a job listing by Curse looking for a Python programmer.

not that I would aim for a job in the layoff factory that is game design

49cf0d No.1342


df1f36 No.1385

>>1196
>Implying algebraic data types aren't objects without inheritance

df1f36 No.1386

>>1311
>bourne shell is a comfy language
It's comfy up until you have to do anything over one line. Then it's just a mess.

a07bd4 No.1424

>>1082
using goto is perfectly fine.
it's faster than deep recursion because muh stack and is more readable than tail recursion
>fite me

a07bd4 No.1425

>>1126
>uses sed to replace python with perl
do even you perl?

d21557 No.1436

>>1135
Haskell

d21557 No.1437

>>1424
How's it more readable than tail recursion? Are you a dog?

a07bd4 No.1438

>>1437
because it uses simpler principles than recursion. recursion is rarely ever needed or even useful. also, even though i didn't think about it, it's also faster than tail recursion. though your compiler will probably convert it to that for you. or possible a while loop. either one(while or goto) should have the same performance even on a non-optimized compiler(let's remember why those had to be created in the first place)
that being said, goto is master race to recursion.

d21557 No.1441

>>1438
Yes, in tail recursion is heavily optimized in any compiler worth looking at for a functional language. However, I don't see how it's more readable necessarily, as I honestly feel a more functional style is more readable and abstractable than an imperative style (though the latter has its merits, and in languages like Haskell you can still do imperative programming without any trouble, but for the sake of the `goto' discussion it is my personal opinion that a recursive implementation is more readable in most cases (where a recursive implementation matches the scenario in structure) and even when a recursive implementation does not mirror a problem's structure, a higher-level control flow is still preferable to a goto-based one, as data flows through a higher-level control flow more fluidly and elegantly).

d21557 No.1443

>>1438
I am >>1441. The fact that goto uses simpler principles than recursion does not make it more readable. Why don't you just forego any higher-level language and get a punchcard-based computer?

You said that recursion is rarely ever useful or needed, which is blatant bullshit. Recursion is the most natural way to model construction and manipulation of nested data structures like binary/n-trees, lists, tuples, etc. and allows for heavy abstraction and concise code. Imperative styles (including goto-based ones) cannot match recursive styles in effectiveness, readability, or conciseness in such areas.

0b65ed No.1460

>>1424
You can't do modular iteration with loops, and if you only have goto you have to do a bunch of manual work versus passing parameters to tail called routines.

000000 No.1508

1. I've been doing Java and J2EE for fifteen years, and I still think it's a great solution for large webapps.

2. I don't think we should expend the effort to make things easier for people to get into software development. The people that are complaining about this are consistently people we don't want.

813bdc No.1549

>>1508
That id
nice

8c5ea3 No.1552

goto statements are great

b57f8d No.1554

Java is actually pretty decent.

f14a7a No.1615

>>1298
C++ is better than C

bd9e6f No.1645

mips is more scalable & more efficient than arm

827a58 No.1675

A text editor should not have "built-in support for GnuTLS, GTK+ 3, ImageMagick, SELinux, and Libxml2".

Is Emacs *actually* an operating system? I really, really thought it was supposed to be a text editor.

6d743e No.1701

>>1549
it means he is using tor

62e12a No.1706

>>1552
I totally agree. I don't trust my compiler to optimize extraneous function calls away so I do it myself. They are also very necessary for multi-level breaks and continues, and state machines (ncurses menus is a good example).

Lua is the best interpreted language that exists. Python is slower, doesn't interface with C code as well, and is bulkier and harder to learn. The standard libraries in Lua are small, but there are plenty of libraries written for it to extend its functionality to just as much as Python's. And writing those libraries is easier to proof-of-concepot because the language is simpler and its easier to optimize because you can rewrite in C and maintain the same library API.

63c6b3 No.1713

File: 1426561685163.jpg (28.11 KB, 499x327, 499:327, java problemfactory.jpg)

1. That CSS is so bad that we should start over with something else.

2. XHTML should still be a thing, because if you are too fucking stupid to match tags, stop making websites.

cd34b9 No.1749

>>1713
>CSS is so bad that we should start over with something else.

that or just let the indians have it

65a2ae No.1795

C ought to be good enough for anybody.

OOP and *most* other paradigms of recent adoption impair more than they help.

Non-strict evaluation is, most of the time, unneeded and should not be default.

The web is a shanty town that somehow is perceived as an ivory tower.

The UNIX philsophy still has a lot to offer.

29778c No.1805

C is obsolete and any large piece of software written in it is prone to more bugs than a stray dog and more leaks than a screen door on a submarine. The only way to get around that is to have someone that both very much knows what they are doing and have them spend a lot of time on it. Time and skill that could have been more better suited for a more developed tool.

I think Lisp is a strictly academic language made for compiler and AI research and has no practical purpose outside of those fields. It kind of works as an extension language, but I think other languages would do just as well but have actually readable syntax.

There's no good scripting language.

There's no good tiling window manager.

Emacs is the best operating system.

BSD is pointless and only serves to give free software to corporations.

>>1713
These are unpopular?

639234 No.1820

Langs like Opal are the future of programming.

afa50d No.1822

>>1713
I don't have many problems with CSS. It's true it could use some better syntax to define whether the rule affects a tag only of it is inside a specific kind of tag and the likes, but it's pretty straightforward.

You NEED a search engine to completely grasp it thanks to the retarded amount of attributes out there, but it's understandable due to all the configuration options it has.

The only thing I hate about CSS is how each browser tries to do stuff its own way, but that's not really CSS' fault. What would you change about it?

I completely agree with the XHTML thing, though. Have you seen HTML5's rules doc? It fucking froze my browser while loading. Why couldn't XHTML be a thing?

afa50d No.1823

>>1438
>Recursion is bloat
Funniest thing I have heard today. Thanks, Uriel.

2bb13b No.1824

>>1438
>>1443

>You said that recursion is rarely ever useful or needed, which is blatant bullshit. abstraction and concise code.


Seriously wtf. Has anon never used a fucking file browser or what?

7716ba No.1829

PHP is the fukken shit

ee2ff6 No.1830

>>1823
He's right, a recursive function will blow the stack compared to a regular looping construct.

d35d85 No.1868

>>1830
>What is tail call optimization

e85640 No.1873

>>1868 different anon here. idk what you're trying to do here but you better be trolling.
recursion combines all that is evil.
it's harder to understand than a mess of gotos and exceptions, slower than letting a three year old calculate by hand and probably has all sorts of other issues.
some optimizations are not an option in debug mode either.

827a58 No.1876

>>1873
>it's harder to understand than a mess of gotos and exceptions
Not terribly, especially when applied to tree-like data structures or parsing some language (i.e. things it is meant to do).

>slower than letting a three year old calculate by hand

That "tail call optimization" the other anon mentioned is essentially inlining the recursive function; it reduces the function call to a loop like any other.

>hand and probably has all sorts of other issues

This is wild speculation at best.

>some optimizations are not an option in debug mode either.

This depends on your language. C, for example, often wrongly assumes the programmer knows the right thing to do - "if you didn't want another stack frame, you should have restructured your entire code base".

On the other hand, some functional languages, like Scheme, depend on recursion. A Scheme implementation that didn't optimize tail calls would be non-conformant.

41f57f No.1877

if a given language only has one numeric type its automatically shit

ad0f1e No.1891

I hate C++. It is worse than Java.

I like C and JavaScript.

d35d85 No.1893

>>1891
>I like C and JavaScript.
Opinion discarded.

603c54 No.1897

>>1893
i like lua & the D

0cf75c No.1914

Even though javascript is fucking ridiculous sometimes, it's fun.

7418ae No.1917

>>1195
kill yourself

efb89b No.1920

The .net platform is objectively the best platform for rapid software development.

C# > Java

1f2ff6 No.2033

>>1086
Stallman has extremely autistic tastes, preferences, and personal opinions
>only listening to Gregorian chant when coding


That said he's done a great favor to the software world. I appreciate Stallman, but my opinions reflect that of Torvald

>>1891
>liking java
>liking javascript
inhale helium

0d5c47 No.2036

>>2033
Is it at least OK to like VM's? Java would be so cool if it wasn't Java.

dea597 No.2041

inheritance is rarely the right answer

827a58 No.2046

>>2036
>Is it at least OK to like VM's?

Sure. All of the cool operating systems are doing it now.
>https://en.wikipedia.org/wiki/Inferno_%28operating_system%29

9a0fbb No.2047

>>2036
Use Scala.

9e79d5 No.2050

>>2036
No. VMs are why java is bad. It compiles to C. Why not be a heterosexual and just code in C?

9a0fbb No.2052

>>2050
Do you have any idea of what are you talking about or are you just spewing le ebin /g/ maymay?

d7e5e8 No.2082

>>1082
Richard Stallman is a fat kike who tries to involve himself with Linux even though everyone say he turns people away with his "free software" movement

73835c No.2129

>>1082
>I think Python is the best general purpose language if speed doesn't matters.
That's a pointless opinion because speed always matters.

827a58 No.2133

>>2129
>That's a pointless opinion because speed always matters.

"Matters" is subjective.

For example, Python definitely has too much overhead for finding huge prime numbers - it would take months to do what would take a well-optimized Fortran program weeks or days.

On the other hand, Python is more than fast enough to handle making little dialog boxes, and it can do it more succinctly than C or assembly ever could.

4ec440 No.2137

File: 1429383957238.gif (1.71 MB, 245x210, 7:6, 1418271855338.gif)

>>1082

C# and the .NET platform are objectively shit tier

MS is pandering to developers by making these platforms open source because it realizes how irrelevant its OS is becoming on the desktop market

827a58 No.2141

>>2137
>because it realizes how irrelevant its OS is becoming on the desktop market

Seconded. Open sourcing .NET just screams "Embrace, Extend, Extinguish".

4dfe3e No.2220

>>1795

>The web is a shanty town that somehow is perceived as an ivory tower.

I agree wholeheartedly. This shit needs to be fixed.


4ec440 No.2242

File: 1429888119101.jpg (78.82 KB, 441x403, 441:403, varg.jpg)

>>1178

>>1795

>oop is shit

>OOP and *most* other paradigms of recent adoption impair more than they help.

These guys know the score.


1c12ed No.3202

>>1920

Even without .net C# is a far superior language. Actually, it's better without having being forced to use Microsoft through .net


3c879f No.3205

>>2242

Thirding that OOP is pure shit.

Extreme complexity. OOP's popularity within large companies is due to "large (and frequently changing) groups of mediocre programmers." according to Paul Graham, the discipline imposed by OOP prevents any one programmer from "doing too much damage." and from figuring out too much.


402373 No.3233

>>3205

Paul used to be my idol, until I realized he's basically a rich version of the TempleOS guy.

Arc has a been WIP for 10+ years. There's no excuse for that


b34e50 No.3251

java is a great language, but maven is retarded and needlessly complicated

javascript is for babbies

perl > python


3b43ee No.3313

>>1425

>Not calling grep in your perl script and using the substitute command


50d1a8 No.3316

>>2082

I fully agree with you. I don't like RMS either, he's done more damage than he has good.


1103dd No.3317

>>1438

>recursion is rarely ever needed or even useful

RRREEEEEEEEEEEEEEEEEEEEEEEEEEEE


1103dd No.3318

>>1675

What are the responsibilities of PID 0?

How hard would it be to patch Emacs to take care of those?


1103dd No.3319

>>1805

>There's no good tiling window manager.

What's wrong with i3?

The only problems I have are when programs decide that the best way to update a window is to close it and open a new window before the user notices.

Seriously, what the fuck is that about GIMP?


1103dd No.3320

>>1873

>it's harder to understand than a mess of gotos and exceptions

Are you one of those people that thinks mathematical induction doesn't actually work?


bf176c No.3322

>>1920

>C# > Java

This is a thread about unpopular opinions, not wrong opinions.


1103dd No.3323

>>3322

>>1920

I've never touched C#, could you summarize it for a C/C++ guy?


495b7a No.3332

The future will be purely based on math.

>>3323

less shitty than java, more features than C++.


3de195 No.3333

Programmers in general are pretty stupid. Every time a new fad comes along, everybody adopts it without thinking about whether it's good or not. There's no critical thinking whatsoever.

Consider all of the crap that programmers have to put up with today, like the stigma against goto, OOP everywhere, the stigma against macros, the use of exceptions instead of conditionals, slow interpreted languages everywhere, "solve this in 5 minutes, if you can loser" coding puzzles in interviews, unproductive "agile" project management everywhere, and the current diversity/SJW trend. There wouldn't have been any of that if somebody stepped up and said "hey guys, this is fucking stupid and you're stupid for liking it" back when this stuff was first being adopted. Instead, everybody mindlessly jumped on the bandwagon, and now the future generations have to suffer for it.


495b7a No.3337

>>3333

i… i think i love you.


67c7af No.3338

>>3333

You're largely right, except

> like the stigma against goto

Goto is awful, and breaks down the structure of programs into awful spaghetti code.

If you're using goto except as a glorified multi-level break, (or some crazy C optimization if you actually have the aptitude) you are a shit coder. 99% of the use of goto is damaging, and the stigma is warranted.

> the stigma against macros

Mocros are useful, but best avoided where possible. They are great for reducing clutter, code generation, and improving performance, but are also a level separate from the language and therefore a source for confusion, error, and yet another syntax dancing in your program (one that is not aware of your programs types or other logic, morever).

> the use of exceptions instead of conditionals

As a programmer, I hate exceptions, but as a project lead, I love them. They are shit as far as performance, logic, and standards go, but wonderful when you have an interface or library and you want to absolutely make it shit on anybody who tries to misuse it. When you have people you don't trust to not lazily forget to check output, exceptions force them to do it.

I agree with everything else, specifically the SJW shit, slow interpreted languages, and "agile". Being a programmer is pain around all the stupid fucking fads everywhere. I came in to my job and everybody was stoked about this big new Rails project they were working on, and 4 years later, we have this big, unwieldy clusterfuck that runs like goddamn garbage desipite all attempts at optimizations. We have only a few hundred people using this product regularly, and still have to have 10 servers running the application in parallel just to handle that.

OOP is only really useful for physics simulations, video games, and easy, fast dynamic dispatch and should be used nowhere else, though it is a slightly more elegant solution than duck typing.


495b7a No.3340

>>3338

>exceptions force them to do it

What the fuck is a Optional or Either type, anon?


67c7af No.3341

>>3340

Optional/Either types are great solutions, but they aren't available as a language feature in all languages, they are difficult to work into the frameworks of existing projects that don't already use them, and they still have the same problem of being ignorable by bad/lazy/tired developers. It's a lot easier to modify an existing library or framework to throw exceptions at bad use than to change the interface to output homemade container types that can still just be ignored. Exceptions still have the advantage of crashing the program or at least kicking out to the innermost catch if the programmer fucks up. The only way to abuse exceptions there is if you use an empty catch or something of the sort, and that's only done by outright programmer incompetence.

I probably should have mentioned that I primarily work with C++.


495b7a No.3342

>>3341

I suggest you a) get better programmers and b) use a better language c) stop being a massive cuck


67c7af No.3344

>>3342

a) I don't do the hiring, b) I really enjoy C++, c) sorry for having a job, bro


4af06b No.3354

XSLT is great


9a6ff5 No.3387

All I ever needed is ASM/C/C++, most of my programs are some mix of those three, using the best from each one.

inb4 suicide suggestions


495b7a No.3388

>>3387

I have never thought that C++'s feature and bloat is useful.


9a6ff5 No.3389

>>3388

they are, for prototyping :^)


fcdc28 No.3500

Software development methodologies should die slowly in a fire.

Not management, not budgets, not time schedules, development should be the focus.

Imagine medical or automotive software constrained by the above factors.

Would kill myself in my new fancy robot car going full retard KIT2000 in the pedestrian zone if this will be the future.


fb4f11 No.3757

It does not annoy me when normalfags are amazed at what I am doing, yet.


47ca45 No.3762

>>1178

inheritance is shit, not objects


98cea4 No.3770

File: 1451186836856.gif (217 KB, 173x261, 173:261, 2b6.gif)

So, according to this thread this is unpopular.

I like python and think Guido is a cool guy.


44feeb No.3918

>>2033

Does Stallman still code? B-)

The problem with him is not that he's autist. It's that he thinks that the whole world should live according to his impossible standards. Case in point: For years he lived on MIT, rent-free. That maladjusted him into thinking that everything is for freedomz, or something.

>>2141

>Seconded. Open sourcing .NET just screams "Embrace, Extend, Extinguish".

So open source wasn't the panacea? B-)


3102a3 No.3929

>>1137

We've got a lisper over here


44feeb No.3933

File: 1455182634978.gif (176.33 KB, 363x511, 363:511, Karl Marx.gif)

>>3918

To further develop my thoughts.

The reason that programmers are treated as beta plebs is because of Stallman. His idolatry of programming, something that *should* be made out of *kindness* has done unimaginable damage. His ifs and buts on making money on programming renders it effectively a non-commodity.

And if programming is not a commodity, then it's nothing to negotiate about. And big companies just loves workers who don't negotiate. Even if it will hurt them in the long run. Such as EA's etc policies who made "crunches" and "death marches" standard operating procedure. No wonder that programmers leaves the gaming industry and goes to finance or oil when they hit the raising a family-age.

I'm not saying that this policy was only possible with Stallman. I'm neither saying that this policy was predetermined. EA could have not sucked Origin dry and so on. And I'm not saying that Stallman is actively obstructing the unionizing of programmers. But still, Le Mememan haven't done shit for 30 years and farts spooks about the nature of programming is EA's best enemy.

If you got to have an idol with wild hair and beard, at least do it properly.




[Return][Go to top][Catalog][Post a Reply]
Delete Post [ ]
[]
[ home / board list / faq / random / create / bans / search / manage / irc ] [ ]