/tech/ - Technology★

The fuck are you still doing on 8chan? Don't you know the internet is harmful?

>>992256

Is there any alternative?

>>992261

Wearing le anonymoose mask and meeting people IRL with voice changers.

All technology is harmful, so the other option is real life, but real life is spyware so you have to mask your face and voice to stay anonymous, and carry an automatic rifle if you want to stay safe from peaceful refugees, but it's illegal to carry self defense tools so you have to do it at night.

>>992251 (OP)

>taking advice on what is harmful

>from an hero

sage

>>992287

Bill Clinton?

>>992293

Some stuff like the compiler is just retarded but it's more common sense tbh.

Obviously java is very fucking bloated.

Seriously? Both Suckless and Cat-V have a lot in common, mainly in terms of being wrong. Seriously, the single argument against Clang is "it's written in C++". If you're sane enough, you should know that you can write some pretty sane code using C++ aswell as write completly unreadable spaghetti code in C. There are a lot more reasons to not agree with them, though single points there are actually nice (eg. systemd)

>>992297

Agreed anon.

>>992251 (OP)

Most of the stuff you can read on cat-v/suckless is about personal preferences, that don't really change the outcome. If you like writing in C, you don't really have to switch to C++ (ofc if not doing a lot of OOP programming). If you're okay with C++, there's no reason to switch to C (except occasional embedded/OS programming).

Heck, the only things that are actually bad, while being mentioned on both sites are:

>systemd

>some gcc extensions

Why so? Because they actually disallow you to use alternative software, according to your preference. I mean, GCC is to some extent okay, though Clang is better due to having a lot smaller architecture-specific code amount, resulting in even smaller attack surface. So, in the end, actually only systemd argument is a valid one, but that one is commonly accepted almost everywhere (except plebbit).

>>992302

Go doesn't have anything to do with gender.

>>992310

Great news! I'm rolling in queers. Where can I buy go tutorial?

brownish right-wing ceterofluid cuckold

>>992311

Who is it? Ken Thompson?

>>992315

Rob Pike, I think.

>>992251 (OP)

Everything on this site implies that UNIX is the right way.

Also, check 'em.

>>992329

And there is only one REAL UNIX.

https://www.youtube.com/watch?&v=E7q7O4Hok6o

>Wants to talk shit about java

>Uses quotes

you cant fuck up something that easy.

C++ is too arcane and bloated for SJW snowflakes to ever learn.

>>992310

>sees the pic related

this reminds me that the word "gender" does NOT refer to the real biological sex, but rather your own opinion of it. the word "gender" is to be considered harmful and SJW faggotry.

>>992315

I seriously hope you are trolling.

>>992315

Maybe it's Dennis Ritchie.

>>992293

Everything on both the "harmful" and "less harmful" lists is the same C/UNIX bullshit. It's "use this C library instead of this other one" instead of actually saying anything that sucks about the standard C library itself. Anything that recommends awk and ed sucks. There were better programming languages and editors around in the 70s. The rest of that list is just as bad.

>>992304

>"*Gnome seems to be developed by interface nazis, where consistently the excuse for not doing something is not "it's too complicated to do," but "it would confuse users"." - Linus Torvalds

If only that were true about GNOME, but it's not. Not confusing users isn't a reason not to do something, but it is a reason to do it right. "It's too complicated to do" is not a valid excuse either if other people 50 years ago who didn't have 15,600 "programmers" could do it.

>Object Oriented Programming is Inherently Harmful

Users of Smalltalk, Common Lisp, Simula, and other non-UNIX OO languages nearly all said that OOP was a good thing. The whole craze of adding OOP to another language started because of the benefits these OO programmers actually had. CLOS was approved in 1988, so if your OO language in 2018 isn't as good as CLOS was in 1988, it sucks, at least when it comes to OOP.

https://www.cs.cmu.edu/Groups/AI/html/cltl/clm/node260.html

>Symlinks are an abomination

Like everything else in UNIX, it's the UNIX implementation that sucks. Symlinks were not an original part of UNIX, so all the "tools" had to be hacked to support them.

>>992307

C++ is a UNIX language created at AT&T Bell Labs, just like C and the rest of UNIX. Using C++ helps C. Using C helps C++. The bullshit C vs C++ "war" is no war at all. It's the same people, the same company, the same compilers, and so on.

Date: Mon, 4 Nov 91 13:41:13 EST
Subject: C++ Is Good For Fertilizer

This message is in reply to a "discussion" spurred by the
"observation" that (I quote roughly and from memory)
"Computers are used by 50% of the people in the world, and C
is used by 50% of all programs, therefore improving C [ie
incrementing it and discarding the result] will potentially
make the world a 25% better place to live."  I am not making
this up.

The original subject line for the discussion was "C++ is
good for the world."

>>992454

>it's the UNIX implementation that sucks.

Please explain why and give a solution.

Anything come up by a community is inherently wrong because the majority of people are incapable of making correct conclusions and reasoning and the ultimate word of a community comes from the majority of its members.

>>992561

>glibc doesn't play nice with static linking

That defeats the whole purpose of having a library. It makes as much sense as statically linking your kernel and drivers with every single program. On some OSes, like Multics, ring 0 code also resides in segments and is accessed through call gates.

>I believe their rationale is that these supposed benefits aren't benefits at all and do not help programmers;

You might be right if you define "help programmers" as requiring 15,600 programmers to do the job 1 or 2 were capable of doing in the 80s, but that assumes these 15,598 programmers wouldn't be doing something more productive if they didn't have to worry about 15 million line kernels and bugs that typical 60s compilers could catch.

>consequently, they're correct as OOP/OOAD does not help, doesn't do what is often claimed, and is responsible for horrible software quality.

The entire reason for OO hype is that these languages did do what they claimed and users wanted those benefits for languages they were using. It was OOP users who said that C++ was bad OOP. Now, C++ is the basic idea of OOP for most people, so they say OOP sucks instead of C++ sucks.

>Additionally, just because some view it to be beneficial or a net positive, doesn't mean they're right or that it actually is.

It's a benefit because it improves code reuse enormously. Code reuse doesn't only mean not having to copy and paste, it also means sharing machine code and libraries.

>I really wish you would provide more context for these statements; operating system composition is an incredibly engaging topic that only remains so as long as particulars are given.

A symbolic link is just a file containing the name of another file. It's the "tools" that are broken.

>It's truly amazing to read that you believe the procedural and OOP paradigms are one and the same.

Where did I say that? C and C++ both have the same problems and share most of the same core. C++ includes the C standard library. If you use C++, you will not be getting away from C's serious flaws and misdesigns.

>>992562

>Please explain why and give a solution.

Because the "tools" weren't designed to work with symbolic links from the beginning because they were a good idea from another OS that UNIX weenies poorly hacked into UNIX. The solution is to replace all UNIX "tools" that handle symlinks.

   It's really sobering to think we live in a society that
   allows the people who design systems like xauth to vote,
   drive cars, own firearms and reproduce.

I'm just graduating from business school*, and was
interviewing with a consulting company that does full scale
analysis, design, and implementation of information
technology systems for companies.

My interviewer looked at me with a rather puzzled, sad
expression on his face, and asked mournfully: "We put money
into Unix.  We put a LOT of money into Unix.  *Why* isn't it
any turning out to be any good for doing really useful
projects?"

We decided the answer was obvious.


* [At business school, I've mainly learned that business is
* set up about as sensibly as the X authorization file.
* *Sigh* So much for $70,000.]

>>992251 (OP)

Reminds me of the cancer labels on everything because of California and there may be a trace of an iota of cross contamination from a facility across the street that handles something that may or may not be carcinogenic. Eventually people stop taking the warnings seriously and the things that are actually harmful aren't treated with scrutiny.

>>993318

>>992454

Daily reminder that Lispfags couldn't even write a simple text editor, yet they think they're in position to discuss OS design

>>993436

The more I argued with them, the better I came to know their dialectic. First they counted on the stupidity of their adversary, and then, when there was no other way out, they themselves simply played stupid. If all this didn't help, they pretended not to understand, or, if challenged, they changed the subject in a hurry, quoted platitudes which, if you accepted them, they immediately related to entirely different matters, and then, if again attacked, gave ground and pretended not to know exactly what you were talking about. Whenever you tried to attack one of these apostles, your hand closed on a jelly-like slime which divided up and poured through your fingers, but in the next moment collected again. But if you really struck one of these fellows so telling a blow that, observed by the audience, he couldn't help but agree, and if you believed that this had taken you at least one step forward, your amazement was great the next day. The unix weenie had not the slightest recollection of the day before, he rattled off his same old nonsense as though nothing at all had happened, and, if indignantly challenged, affected amazement; he couldn't remember a thing, except that he had proved the correctness of his assertions the previous day.

>>952496

>>992311

Somebody needs to get this guy off the soylent and quickly!

>>992251 (OP)

>shilling this hard for plan 9

This is one retarded list of harmful things, boyo. Plan 9 been in development for 100 years and still only has 22 users. Just fucking stop.

>>993466

cat-v isn't about plan9, an OS on which development has been halted for maybe a decade

>>992256

THE INTERNET IS SERIOUS BUISNESS

>>993451

DELETE THIS

>>993453

Pike has been a weirdo since AT&T started Plan 9

>>992302

>Why are they sucking Go's gendre non-binary's dick.

Because it was made by Ken Thompson and Ken Thomson is God.

>>993318

The purpose of the C standard library is to provide functions that wrap around the system calls. Dynamically linking it effectively means you put system calls behind two call gates instead of just one for no good reason.

It makes sense to put system calls behind call gates. If you print a string to stdout that might mean physically writing to disk, or printing it on the screen, or any number of other things. Libraries like OpenGL/Vulkan, or any other library that does completely different things on different machines should be dynamically linked, but most shouldn't.

>A symbolic link is just a file containing the name of another file.

Cool. Is there any reason for it to exist? Other than the UNIX paradigm being broken?

>>997534

>Cool. Is there any reason for it to exist?

Are you retarded? How about a copy of a file that doesn't take as much disk space. How about organizing files without touching the original (what Windows plebeians do by hand in their "torrent renamers"). How about being able to easily change /bin/sh with a symlink?

Symlinks are good, period, what's bad is their inconsistent handling by various tools.

>>997534

>Is there any reason for it to exist?

Plenty, if done properly they allow for incredible flexibility.

>>997539

>what Windows plebeians do by hand

You know windows has symlinks, right?

>>997554

Of course I know, but the typical luser will never use the command line and the GUI only allows you to (tediously) make one symlink.

retard threads considered harmful

>>992454

and what is wrong with awk?

>>997539

>How about a copy of a file that doesn't take as much disk space.

A hard link serves this purpose much better. You wouldn't want a "copy" to suddenly stopped working when the original gets moved/destroyed.

>How about organizing files without touching the original (what Windows plebeians do by hand in their "torrent renamers").

What does that mean? Also searching for "torrent renamers" returns nothing relevant (a program for batch renaming of files on Windows and information on how to edit .torrent files).

>How about being able to easily change /bin/sh with a symlink?

You almost had a good point but fucked it up. /bin/sh can be a hard link to your favorite shell. Now what you could have said is stuff like /bin being a symlink to /usr/bin etc. However, this is just another example of the retarded UNIX paradigm. Programs shouldn't have paths hardcoded into them (except maybe ONE configuration file in /etc), and the environment should die.

>>992251 (OP)

> Will I became cool minimalist hacker then?

You're probably better off to learn Forth and forget all that C/Unix stuff.

>>997558

>but the typical luser

Will never encounter a case in which they would benefit from manually creating symlinks, besides shortcuts.

>>997876

I just gave one, m8. Rename function for bloated torrent clients is almost needed by the average user.

>>997834

You don't seem to understand the point of symlinks. First, they're a lot more simply to understand/less opaque than hardlinks.

>A hard link serves this purpose much better. You wouldn't want a "copy" to suddenly stopped working when the original gets moved/destroyed.

Yeah, I used the wrong term. I want more than one name associated to the same file; yes, I know that hardlinks can do that too, but I like being able to see what points where.

>What does that mean? Also searching for "torrent renamers" returns nothing relevant (a program for batch renaming of files on Windows and information on how to edit .torrent files).

Being able to seed a torrent while renaming the downloaded content. The torrent client implement a translation db instead of using links, I think.

>You almost had a good point but fucked it up. /bin/sh can be a hard link to your favorite shell. Now what you could have said is stuff like /bin being a symlink to /usr/bin etc. However, this is just another example of the retarded UNIX paradigm. Programs shouldn't have paths hardcoded into them (except maybe ONE configuration file in /etc), and the environment should die.

The UNIX file hierarchy is bad and outdated, but the concept of having all the binaries, configuration file, etc... in one place is a good idea. What would be an even better idea would be to have a new directory giving you the windows way (grouping stuff by the package/application it belongs to) via links.

Now, I don't get what your position is? Do you think symlink and hardlinks overlap too much? I think so too. I also think that something cleaner, like plan9's bind could be better. But as I said, my main problem with symlinks is the inconsistent handlings by the *utils.

>>997887

>The UNIX file hierarchy is bad and outdated, but the concept of having all the binaries, configuration file, etc... in one place is a good idea.

Perhaps, but having a convention like the name of the package also being the name of the executable if any (/bin/awk becomes /apps/awk/awk) would be just as good if not better than Windows' symlink-heavy method.

>Now, I don't get what your position is?

In an ideal world we would have figured out a better way of building a filesystem. Symlinks are a bandage over the underlying problems with our current ones. The file system we have on Unix was created for a simpler time: back when the vast majority of programs did not have to worry about state or configuration. Ryan Dahl wrote a pretty good bit about the Unix vision vs what we have now (can be found here: http://harmful.cat-v.org/software/node.js original post was deleted). We need to figure out solutions to the problems of overwhelming complexity, ever increasing state etc. which is something the cat-v people realised. Unfortunately their solution is Plan 9: trying to go back (which is not something you can do in tech) to the old ideas of Unix, while also making everything distributed (even though literally no one aside from Rom Pike cares about distributed computing).

Until we get something better, we are going to have to deal with a lot of shit. One of these pieces of shit will be the way we structure filesystems. And symlinks are just one of the fixes which actually turn out to be a pain in the ass but you still have to use because there is, in many cases, no way for you to make software behave otherwise. And if you think you can fix the situation by writing better utils please do, you would be doing everyone a great favor.