ITT: unpopular opinions
If you need a debugger, you should not be writing software. It's like a terrible mason putting up skewed walls and then using a "wall straightener" to fix it.
>>51949146
but if the end product is the same then why does it matter
I think software is a little more complicated than laying bricks.
Debuggers are nice, but you shouldn't rely on one since some languages don't have debuggers. What you should be relying on is well made and thought out test suites.
t. Rob Pike
>>51949146
Could you explain your reasoning more?
From my point of view a debugger is more like building a mock-up, something you can use to see what things should look like vs what they are now.
I think patching would fall more into the idea of a "wall straightener" as it fixes issues you already shipped but software is fucked as is and I doubt that will ever go away.
>>51949146
I like debuggers, but not for building software. More for ripping software apart.
>>51949184
>>51949181
>>51949175
Code should be readable and understandable as is
If that is not possible, either the code or the programmer is bad.
>>51949201
>Code should be readable and understandable as is
I agree with you here but that dream is dead, software is touched by too many people to be sure exactly what is does anymore.
Just look at that recent CVE Juniper published, they pushed code for products they didn't event know was shipping with own'd code.
>>51949155
it usually isn't.
>>51949146
Nice bait, senpai.
Can we see your code you wrote without a debugger?
>>51951387
check out the linux kernel
>>51949201
You realize that companies use multiple people to program, right? So you don't know all the code, that's the reason you need debugger. If you change something there is a chance that your change can interfere with other parts of the code.
When you stop being a script-kiddie maybe you will understand
>>51951431
what the fuck does a debugger do for you in the situation you're changing code like an idiot?
Are we taking boundary scanners or are we talking syntax errors?
Because both are incredibly useful. Especially when working on large projects.
You must be a script kiddie
>>51951464
once you stop reimplementing fizzbuzz in meme languages and actually do something useful you'll start appreciating debuggers
>>51951464
That's why we have Software Engineering and UML/Documentation shit.
But at the same time, the company needs to know if everything is okay with the program.
I have problems with how it is right now, but until someone can come up with a better way of doing it (we have Software Abstractions from Daniel Jackson, but it is still too early to use Alloy or any other more "esoteric" way).
>>51951528
>using SDLC
>>51951490
I'm 5 years into this field and all the people I know with popular and useful projects have all moved away from needing them, including myself. So uh, what? You need a better argument.
>>51951387
Implying it hasn't been fucked in the ass by GDB sum of thousands of years now.
>>51951592
How big is the project you're developing? GitHub shit is not valid.
How many people are working with you?
How critical is your software?
>>51951528
>UML
I haven't seen that thing even once since I graduated and started working.
>>51951415
>>51951611
>>51949146
I think OP likes to work in the dark, turn off all lights and hope for the best.
Why not both? Write as normal and debugger for verification mostly
>>51951633
How big is the company you work for?
Do you use any use-case? Software Engineering? Agile?
>>51951637
>>51951624
holy shit answer questions and stop arguing with yourself.
please tell me what a debugger does for you?
ITT: People that program little programs with few people comparing this shit with Critical Softwares like medical / airplane / etc software that can kill or save thousands of people.
Grow up, kids.
Everyone writes buggy code. Even Linus Torvalds and Theo deRaadt writes buggy code. Grow up.
>>51951663
If there is a bug that can destroy a village because the drone software you created is buggy.
>>51951695
>software that can kill or save thousands of people
>people
>as in humans
most guys on this board wouldn't care about that
>>51951730
>>51951653
currently? around 40 people. yeah, we're doing scrum.
>>51951727
how does a debugger magically find a bug? you think you can't do that without a debugger also?
>>51949146
>I'm going to repeat the same shit that Linus Torvalds spouts in his shitty mail list
>>51951767
I'm not trying to be edgy, just, come on. The average poster on /g/ is quite high on the autismo scale. You think they really can relate to the fact that them writing a buffer overflow caused somewhere immeasurably far away to come to harm?
>>51951727
If we're talking about drone killing more sandniggers than originally intended, that's not a bug, that's a feature.
>>51951415
>the most clean piece-of-shit and the most organized mess in existence
ITT: people that can't even answer what a debugger does for them but they're damn sure they need it and willing to push that belief onto others by arguing in bad faith.
>>51949146
Linux is trash
>>51951663
>What is verification
Oh, look its correct! OP is indeed a faggot!
![[YungLinus].gif](https://i.imgur.com/j0JnwIA.gif "[YungLinus].gif")
>>51949146
Hi Linus ^^
>>51951769
How many people do you think developed the Boeing software that controls the airplane? 40? 400? 4,000?
40 is considered a small-medium company and you know where and when to code, you can talk to the other person. Imagine if there is people all over the world developing the same software? We need to know how data goes and are processed by the System.
>>51951776
You know that are many ways to find bugs, right? Your normal debugger is only a way to do so. Have you ever read Software Abstractions? Formal Methods? Mathematical proofs? There are lots of way to test your software for bugs. When you leave the "safe zone" you talk to me.
>>51951814
Why should I care if people are stupid? Not my concern here.
>>51951819
How you left /pol/? Go back there, and stay there.
>>51951840
kek you can't even write fizzbuzz correctly
>>51951895
>You know that are many ways to find bugs, right? Your normal debugger is only a way to do so. Have you ever read Software Abstractions? Formal Methods? Mathematical proofs? There are lots of way to test your software for bugs. When you leave the "safe zone" you talk to me.
this is actually my argument. debugging is a thing you do. if you think a debugger is neccessary you're really bad at this.
>>51951967
what are you basing this on? your level of butthurt?
>>51949175
Not really
>>51949146
If that existed everyone would use it
Fun fact: Linus Torvalds hates debuggers because he doesn't know how to use them properly (and got a fight with one of his teachers because this)
>intense posterior inflammation and damage control because you need a debugger to understand code
Debuggers are only needed in high level languages because you cant visibly see the connections between expressions in high level syntax because its....high level, abstractions by definition will hide inner workings. People who do bare metal systems will work directly with the outputted assembly, you dont need a debugger with assembly because its immediately obvious where program control is in the syntax
"I don't like debuggers. Never have, probably never will."
— Linus Torvalds
>>51952446
I teach a lot of classes in C++, Java, C#, TDD, XP, Patterns, etc. In those classes I often have the students write code. It is not unusual for me to find a student with his or her nose buried in a debugger, painstakingly stepping from line to line, examining variables, setting breakpoints, and generally wasting time. The bug they are tracking could be found through simple inspection of the code.
I consider debuggers to be a drug -- an addiction. Programmers can get into the horrible habbit of depending on the debugger instead of on their brain. IMHO a debugger is a tool of last resort. Once you have exhausted every other avenue of diagnosis, and have given very careful thought to just rewriting the offending code, *then* you may need a debugger.
>>51951695
God help your clients if you work on airplanes and need a debugger to understand the code, instead of relying on actual software engineering procedures and formal verification.
>>51952486
>I teach a lot of classes in C++, Java, C#, TDD, XP, Patterns, etc.
>I consider debuggers to be a drug -- an addiction.
>typical academic mental masterbation
trying getting a job in industry some time where you are working with hundreds of thousands of LoC
>>51952524
>trying getting a job in industry some time where you are working with hundreds of thousands of LoC
How is a debugger helping you there?
>>51952543
>100k LoC
>How is a debugger helping you there?
those who cant do, teach
>>51952573
So you step through the 100k LOC, inspecting variables? LOL.
stop replying in this thread. direct questions about why a debugger helps are answered with troll remarks without any content whatsoever about the question.
>tfw all these code monkeys and wannabes celebrating all shit that torvalds says without questioning it critically
Wonder why they think that Linux is the pinacle of OS development...
>>51952606
actual industry enterprise quality aerospace
Debuggers are great, as you will never work on a project of any real consequence without interacting with code written by someone else.
>>51952543
>has never actually used a debugger
>>51952446
Absolutely this.
Debuggers are useful in certain scenarios.
When a C program segfaults there isn't really any contextual information given at all but if I load it up in gdb, I can see a stack trace and go about the process of trying to fix what went wrong.
When some big, unwieldy OO JavaScript library I didn't write does something stupid, the debuggers in browsers are invaluable in tracking down what the fuck is actually happening: jumping through a call stack that's 100 functions deep, examining the actual state of the running program, etc.
I mean it's a tool like anything else; it can't and shouldn't do your job as a programmer for you but it can make life a darn sight easier if you learn to use it effectively.
>>51949201
The code being readable has nothing to do with the logic that the code describes, which is where actual difficult to find bugs occur
>>51949146
Serious question here. Have you never:
1. Had to work on code someone else wrote?
2. Written anything in a low(er) level language like C that was more than 1000 lines of code?
3. Gone looking for vulnerabilities in software?
Debuggers don't exist to compensate for sloppy programming. They exist to help humans cope with our own fallibility and the incredible complexity of writing maybe systems. I respect that you put value on adhering to good practices, but not being aware of the kind of things I've mentioned really only speaks to your lack of experience. Not that that's meant to be a put down or anything.
>>51954884
Not "maybe systems", " large systems ". Swipe typing.
>>51949146
>not using every tool available
stay scrub
If you use a application framework then you're a unskilled code monkey.
>>51949146
You don't just start using a debugger after you've finished the code to fix problems after the fact. You use it while you're writing it, from the beginning. Much like a mason uses a plumb Bob and a level throughout the entire process.
You've been doing it wrong. You don't know how to use your tools.
You're a dumbass. When doing large projects, you NEED a debugger because tiny mistakes in codes can throw many problems. You need one to isolate where the problem is.
>>51955624
this desu senpai
I remember reading how in the 90s a software shop at IBM decided as part of a move to literate programming to forbid debuggers, One engineer took to it instantly and quickly became the team's star programmer, magically finding and fixing problems that had befuddled the other engineers.
After a couple of months of this, he let them in on his secret: he had snuck in a renamed copy of gdb and had been surreptitiously using it all along.
The team went back to using debuggers.
>>51955720
>you NEED a debugger
>>51949192
> ripping
Ur cool zero kewl
>>51951464
Regression testing.
>>51949146
>If you need a debugger, you should not be writing software.
https://jsfiddle.net/m66tpc4c/
>>51949201
>t. Rob Pike
When you get a job you will see things are different.
>>51949146
-10/10
>>51957092
btw, open this in chrome/ium
>>51957127
>>51957092
>https://jsfiddle.net/m66tpc4c/
>>51949146
Can confirm. I'm programming with only vim and gcc. Works find and I've learned much more about programming than the guy clicking together a program using a photoshop like IDE.
>>51957190
bugs happen
>>51957657
Why the fuck would you bother bumping a shitposting thread like this?
>>51957864
>Why the fuck would you bother bumping a shitposting thread like this?
>I want my non-tech generals and my consumerism threads instead!
>>51951987
>butthurt
>by your level of retardation
ehehe
>>51949146
Still one of my favourite quotes: "In C I never learned to use the debugger so I used to never make mistakes"
http://queue.acm.org/detail.cfm?id=1531242
K/Q are awesome languages and everyone should learn them
>>51957864
At least this is a different kind of shitposting, instead of your repetitive faggot OS wars, SUICIDE WATCH generals and daily facebook toy shill threads.
I've been programming for a while, the only thing I see debuggers do for me is tell me if something broke, I'm the one that has to look where? Why? And what can I do to solve it?
Usually that involves some research if I haven't encountered that problem before, sometimes I'm able to see the problem before I test it; guess that's what happens when you gain experience.
I get your point if you are talking about shit like intellisense, when I was 14 I used it to program c# and I didn't understand the 'why' of a lot of stuff until I 3 years later that I started learning c++ in codeblocks and getting some big ass books.
IMO debuggers are great tools for projects with a lot of lines of code, deadlines and your need for a paycheck (I've never worked in large programming teams, I guess they can help there)
I'm still learning and just recently got into college and my method has been:
- design the project in paper
- code it
- test it until you can't make it fail
- publish it
>>If you need a debugger, you should not be writing software
Uh, what? Are you even tangentially aware of anything to do with software development?
>>It's like a terrible mason putting up skewed walls and then using a "wall straightener" to fix it.
It's called a square. It's a tool. Used by those in the construction trades to do things like make sure walls are straight. Similar to how a debugger is used.
This is really the stupidest post I've seen on here this month, which is really saying something since the kids are all out of school and shit posting for our own little repeat of the Eternal September type scenario.
>>51949146
You're out of touch and naive, really.
You have no idea how much research money is poured in code verification.
Both formally, which is fundamentally where research in automated axiom provers is applied.
But mostly through smart "simulations" and trying to verify code in probabilistic sense.
I mean crazy money, for crazy hard problems, when you're dealing with millions of lines of code that has to be rock stable.
And this is "debugging" the logic of the code, otherwise the code is perfectly valid and compiles.
So, you tell me, when trying to debug in this higher level, do you afford to fuck with misusing brackets or invalid arguments and whatnot? This shit is taken for granted, it's like tying to do algebra without knowing your multiplication tables.
>... and then using a wall straightener to fix it
And this is where you go retarded.
Show me a standing building or structure that was build without levers and similar tools. Or any precise mechanical structure not using jinxes to keep shit parallel. Or any CNC or any other industrial machine that doesn't have control code to offset errors because "it's built to work right".
See how stupid that sounds?
>>51951769
Scrum isn't really a big deal. I've been forced to use it at uni for a 3 person team. It was useful but we didn't see the point. now I'm at a small engineering firm working on software with 5 other people, it's a nice way to see what's happening, owner likes it allot.
Scrum isn't a shit look how badass we are. It's just an organisation tool. It's pretty basic but does what it needs to do.
Also a team of like 10 or less is fairly small. Especially when you know everyone and can annoy them on what they changed or implemented.
>>51952486
I can understand where your coming from with a uni project, where honestly it isn't extremely big. When you write all the software, you should understand the possible failure points and can inspect it by eye.
But honestly, having 50+ people working on the same code, with the majority of the code written before you started working at a business makes a debugger an essential tool to see where the data you change is sent to and where it breaks.
>>51952579
Your inability to learn how to use a debugger properly doesn't make them bad.
When I was trying to learn the software I was to be working on. I used the debugger to understand how the code was being called, looking at the stack to see how data was flowing through the code. I mostly had to do this because there was like 10 comments in the 10,000 LOC I had to work with. Debuggers are tools, and all tools have many uses, just depends on your creativity and skills.
>>51949146
I agree
Debuggers are for shitty programmers unsure of what their code is actually doing
But non-autistic plebs will never understand
>>51959153
Meh, like any tool it just makes life easier. You don't need it but it beats the shit out of printing everything out.
