General questions thread

[levi]

One of my dev machines is a P3, and I got Arch installed on that fairly rapidly (the vast majority of time was me scratching my head and running to my work laptop to read the instructions again). It perhaps helps that when installing Arch you have to guide the machine through the installation, formatting discs and connecting to wifi youself for example. There is a helper script which unpacks the CDs worth of packages to a starter base env, but it's up to you to update that to live.

The only thing I've done it it was run the Pandora cross compiler* build to get an x86-32bit build of the tools. My x86 work laptop didn't have enough disc space to build it. In practice, since maxing out the board RAM with 512MB of the stuff and sticking a hard disc that's some 4 times bigger than the one it came with it's become a much more useful machine.

* Not sure which prolific coder gave me this cross compiler, so I can't give props, sorry.

 

[WizardStan]

I would prefer not to have my experiences dismissed like that.

That's why I was very hesitant to say anything, I was afraid it would come off as dismissive. Sorry, I tried to be as gentle as I thought I could be. Your experience is valid, I just don't think you do yourself any good by setting yourself up for failure, nor do you do Linux any justice by repeating the same tired rhetoric that has been rehashed for decades.

As was mentioned, you downloaded something randomly off the internet: that is not the normal expected use of a modern linux system by any stretch, and absolutely shouldn't be conflated with any general difficulties. I'm sorry you had difficulty, but that's not a normal experience, and there was no reason any of us should have known what you were talking about because we didn't have your experience and you didn't take any time to explain it.

Even when it comes to needing to build from source, it's come a long way in even the last 3 or 4 years, thanks to github and configure/cmake. Nowadays if you absolutely need to install something that only exists as source, 95% of the time it's something like "git clone URL; ./configure; make; sudo make install" and the website you get them from will usually tell you these commands if you didn't already know them. And even here, a lot of developers provide x86 binaries you can just unzip and execute.
Linux developers spend so much time and effort solving problems like repositories and standard libraries and tools specifically to make the end user experience better that it virtually breaks my heart whenever someone says something to the effect of "Linux sucks because you need to know how to compile things from command line". Yeah, linux sucks for a lot of reasons, but being able to install and run things isn't one of them (other than the unfortunate caveat of commercial applications that only run on Windows, if you really, really need that very specific program)
[doublepost=1468083213,1468082809][/doublepost]

Got that a lot when I was a beginner, too. Nowadays I sometimes say it myself

Before I tell someone to google it I google it myself. I've been told "google it" many times myself, and when I do it's not uncommon to get hit with pages of people having the same problem with no solution, or vaguely matching problem with solutions that don't work. I can only recall one time where someone told me to "google it" and the solution was actually on there, found only because I didn't know the right words I needed to look for.
Anywho, by the time I verify googling is actually a thing that'd work I've usually found a good solution anyway, so I'll just say something like "10 seconds on google, here's the result with the best answer *click*"
Give a man a fish and he'll eat for a day. Teach a man to fish and he'll starve because it takes time to learn to fish properly. Give a man a PB&J while teaching him to fish and he'll survive until natural causes get him.

 

[Yasu]

Thanks for the kind replies. I think I was out of line.

I like computers and if other people can learn how to use Linux to it's full, why can't I? That is my current stance. And I'm ready to get my misconceptions challanged

 

[tarator]

I'm working 13 years with linux and 8 years without windows anymore. I don't know linux to "it's full". Only the things I need. But now I mostly use it for work and I'm hardly fiddling around with the OS itself. I must say: It's a nice OS for getting things done...

 

[sulu]

@comradekingu:
I can make little sense of your wall of text. Therefore I will only answer to some of it.

Yes, with software you have to trust the source code, and the compiler.

Your original objection was this:

The idea with trust isn't to make you feel safer, its to have something to trust.

This is a tautology. It has no argument. Out of good faith I still tried to make my original point clearer:

That still requires you to trust the source code in the end.
Sure, in principle, you can check all of that yourself, but it is unrealistic that any single person knows what every single line of code does, his system was made of. At one point or another you need to blindly trust the work of others.

In reality you have to do this blindly, because as a human being at one point or another your ability to follow a (branching) chain of trust ends. Please don't try to prove me wrong, by assuming a purely theoretical POV, where the whole chain can be evaluated by somebody!

What your argument hinges on, is that free software isn't safe, because it isn't bulletproof.

No. I'm saying it isn't safe, AND it isn't bulletproof, because it's made by imperfect humans. There is a correlation (because both terms mean pretty much the same in this context), but you are the only one reading causality into these two terms.

Which is why your original argument was incoherent. A chain of trust makes sense for every single chain you make, regardless of the state of your core OS.

My original argument isn't incoherent, because your core OS is the very first element in your chain of trust. If that is already broken, then you can't even begin to build a chain.

No, you wrote it. Nice try though.
"Most of Linux' reputation for being a safe system is either outdated (Windows has caught up a lot), or due to the behavior of its users rather than the technology."

To which I said:

"And no, the security of Linux|Gnu is not dependent on what windows does, let that idea stay in the confused mind that wrote it.

I still don't get why you think I would claim that the security of Linux would depend on the state Windows is in.

Even IF statement B made sense, it would still be a non-squiteur argument, because you are suggesting that the reputation of Linux relies on user behaviour, which is a unjust metric for an OS, and also not why the reputation of Linux is what it is.

Reputation is a subjective term. It depends on your perception of the subject. So the behavior of the users of an OS can have an effect on the OS's reputation. It's basic psychology. I don't understand why you don't get that.

 

[Silent-Hunter]

The thing with open siurce software is, if something was wrong with a bit of code, someone would fix it or complain. With open source everyone can see the source code.