UNIX?

Printable View

Show 40 post(s) from this thread on one page

December 7th, 2004, 10:00 AM
poohsuntzu

Quote:

For the hardware my argument is simple. I simply might not have the internet conection to download a new kernel. This is why I'd prefer to see some sort of linux support on Driver Disks, even if it's just a small file with the code which I need to add to the existing kernel.

The linux kernel and all the modules for it, uncompiled, is around 40-50 megs. Within that kernel, and NOT needing a new one, there are hundreds of thousands of drivers. Like I said, you are still thinking in terms of Windows hardware management. To a Linux kernel, all video cards still use things in the same way execept for minor variations. Those cards that require minor variations have their own seperate code added in that you activate. No download required. If you end up needing a driver in the future (chances are, not) then you can always activate the driver that already exists in the kernel.

It's hard to explain without having you compile a kernel to see. Have you compiled one before, so you understand what I mean in the difference of Windows "Install a driver per hardware" versus linux "One driver in the kernel can actually run 500 other similar peices of hardware, since they do the same thing under a different brand name"?

edit:

Quote:

Well then people should stop complaining that Windows has got such a big market share. Don't you think so?

No, I honestly don't. Very few people who use Linux, BSD, or Nix complain about window's market share size. OpenSource and Linux isnt here for marketshare. They are here to deliever an operating system that has the capability to be incredibly flexible, moduled, secure, and usable. Marketshare may come, but that's a sideproduct from the origonal goal of Free software and Linux.
December 7th, 2004, 10:13 AM
nightcat

OK. I'll have to do a bit more learning on this. As I said before, I'm new to Linux. I goet the idea, but some questions still remain.

Like:

How does an existing modules will guarantie the future compatibility and an abbility to use all the extra features of the card, or other hardware?

but this is to biger question to answer here, so I'll just sort it out for myself as I get to it.
December 7th, 2004, 10:30 AM
poohsuntzu

The easiest way to compare is like this(and please forgive the formatting example), and keep in mind that this is just an example and thus the basis of it can be expanded upon when discussing other types of hardware.

Problem: "I need a driver for my harddrive that uses DMA 4, SCSI capability, and is a Western Digital."

Windows Solution: Download the drivers for the Western Digital (insert model number here) drive.

Unix Solution: Turn on the options in the kernel that allow DMA and SCSI.

So, instead of having a driver for each harddrive, video card, etc etc etc.. UNIX looks at solving the core of the problem. For another example:

"All hard drives need to be able to read a certain file system. Okay, so we program in the ability that every single hard drive uses... reading. There, we've covered 50,000 different harddrives right off the bat.

Hm, now some of the drives can use DMA. Okay, so let's add in an option for DMA. There, now we have 100,000 harddrives that use the features of -reading- and -DMA capability-.

Ah, SCSI technology. Well, let's code how SCSI works. There, now we have 200,000 harddrives that use the features of -reading- and -DMA capability- and -SCSI capability-."

See what I mean? They solve one problem at a time, and then build on that feature. By doing it this way, you don't need (most of the time) 'per brand name' drivers, because in the end they are all going to access DMA the exact same way. They are all going to use their hardware to read to the drive in the same way. And they all use SCSI in the exact same way. So, it's like building a pyramid. The features that are common across ALL the harddrives are implimented into a single option, covering hundreds of thousands of harddrives.

When a new feature comes along, like SATA, then we just have to turn on the option for SATA and ALL drives that use SATA can now .. well.. use SATA rather than having to make a new driver for each individual drive make/model/brand name.

If/when there are special exceptions (the manufacturer does something NON standard to their hardware (fscking ATI!!!!!!)) then we have to rely on the manufactuer either helping out the opensource community or using their own close-source binanaries.

edit: Granted, when extremely new technology does come out, it will eventually require an upgrade of the kernel because we need to add one new feature for those 20358731486703198 harddrives to use as needed. But by that time, you would want to anyways. Keeping your kernel upgraded, secure, and stable is as essential as keeping Window's core upgraded via Windows Updated.
December 7th, 2004, 10:31 AM
wiskic10_4

Quote:

Well then people should stop complaining that Windows has got such a big market share. Don't you think so?

Not everyone in the world is a computer proffecional. And not even all the profecionals can use Linux. Some of my Uni profesors can't even use Windows properly, and they are considered among the top in their field of IT. Beat that

Who's complaining? I'm not anti-Windows. As a matter of fact, the box I'm typing this very message on dual-boots w/ Win98SE (slightly customized, of course...).

I've recently become the admin of a small network at my University (the computer science laboratory), and hardly consider myself a, as you say, "proffecional" (professional - perhaps your RH box isn't equipped w/ a spell-checker either?). During the break, I intend to setup about 25 out of 30 of my Intel boxes with WIN NT (fully patched and slightly modified, of course). Why? Because most of my colleagues are NOT interested in *NIX systems, nor computer security in general. They are programmers, web designers, and information systems buffs, and are not generally concerned with the "inner-workings" of the particular system or network they are working on... of course, there will be *NIX systems available for those wish to use them (probably equipped w/ Debian, RHFC3, and my favorite, Slackware 10)...

I'm not trying to downplay Windows - Microsoft is very good at what it does - providing a low-maintenance (and insecure, vulnerability-prone) operating system on which things will run easily and automatically (for the next month, anyway, or until the latest patch is installed)...

I'm not of the mindset that Microsoft is "evil" - simple market analysis will show that Microsoft is doing the best it can in releasing "quality" software while remaining a competitive and profitable business - but that's a story for another time...

Anyway, sorry about the rant - I've been drinking (I am Wiski C, ya know).

-Wiski
December 7th, 2004, 01:10 PM
The Grunt

Quote:

Some of my Uni profesors can't even use Windows properly, and they are considered among the top in their field of IT. Beat that

They are obviously not on top of too many people if they can't use windows properly. If you are "on top" of the field in IT, and can't use windows properly, you obviously don't deserve your job.
December 7th, 2004, 01:27 PM
MrNiceGuy

You people are arguing with someone who contends that opiates are a "safe" recreational drug. Why bother?
December 7th, 2004, 01:33 PM
cacosapo

Quote:

Originally posted here by The Grunt
They are obviously not on top of too many people if they can't use windows properly. If you are "on top" of the field in IT, and can't use windows properly, you obviously don't deserve your job.

Besides i cant see a professor that is "on top" on IT but cant use windows -- a professor must know windows to be a "professor", i know a lot of ppl that is "on top" and dont know windows, dont use windows, etc. Example: mainframe experts. They only know windows enough to start terminal emulators and use that comfortable "green screen".

BTW: Most ppl make a confusion about Mainframes: Security x Reliability.

Myth: Mainframes are more secure. In fact, they are pretty easy to hack using the basic approach -- the breachs that all admins leave due to bad administration
True: Mainframes are rock solid in stability. Their Hardware and O.S. were designed to survive to a lot of "situations". Ex. IBM is developing their (mainframe) OS for more than 40 years since TOS/DOS thru Z/OS. Its hard to see an O.S. that can survive to a processor failure (on a MP enviroment of course) or a Memory Bank Failure or a bus failure (that we can find on lower platforms). IBM O.S. can recover from that events.

And the question why everyone go back to MF? Even to the re-centralization of processing (WEB=green screens) Mainframes (due to reliabity) are SO MORE costly that a (low platform) cluster for example.

Imagine them how much costs a Mainframe cluster...
December 7th, 2004, 01:37 PM
MrLinus

Quote:

They are obviously not on top of too many people if they can't use windows properly. If you are "on top" of the field in IT, and can't use windows properly, you obviously don't deserve your job.

Not necessarily. As I type this, I have an article in front of me that talks about a professor at the University of Saskatchewan who has taught engineering and computers (including programming) as well as written many textbooks for over 30 years --- and he's never used a computer.

Keep in mind that often professors are expected to be specialists in a particular area, so they become focussed on that narrow view. Doesn't mean that they will be experts in everything. I know someone who knows Microsoft Win2K/2003 inside-out but he couldn't tell you point 1 about Linux nor could he use Powerpoint beyond the basics. I find that it's not unusual for my programming colleagues to be experts at the language they teach but rather neophyte at basic usage. I suspect this stems from the University method of teaching computers: theoretical mathematic point of view with little to no real world application or hands-on.

Quote:

Besides i cant see a professor that is "on top" on IT but cant use windows -- a professor must know windows to be a "professor", i know a lot of ppl that is "on top" and dont know windows, dont use windows, etc. Example: mainframe experts. They only know windows enough to start terminal emulators and use that comfortable "green screen".

That's a load of crap. No professor needs to know Windows inside out unless they are teaching it. Do they need to know how to get things going? Certainly.

Remember, these are just tools for us to do our work on. Perhaps we should take a lesson from Paul Tremblay (the professor I alluded to above) and try working without...
December 7th, 2004, 01:39 PM
The Grunt

I see what you are saying...

My statement still remains then, if they are this specialized with something, with little to no hands on experience, they aren't experts in my opinion :P
December 7th, 2004, 01:40 PM
chsh

Quote:

Originally posted here by nightcat
For the hardware my argument is simple. I simply might not have the internet conection to download a new kernel. This is why I'd prefer to see some sort of linux support on Driver Disks, even if it's just a small file with the code which I need to add to the existing kernel.

I would honestly like to know what hardware you have (a list of everything) that didn't come with driver support for linux because IME almost everyone includes it.

Quote:

All these small things do make a lot of difference. They are the reason why people will or will not use something.
I already gave an example of Fedora Core2. What, for me, was an interesting task, for a general user might be a reason to stop using Linux.

This is where you're going to have to wake up and realise that you really don't "get it" about Linux. If you have a problem with the way Fedora does things, PICK ANOTHER DISTRIBUTION. This isn't Windows where you only have the current version or two and that's it. There are a load of different Linux distributions out there that do things better/worse, give them a try.

Quote:

Now what's to the services that are run within OS. Linux is also runing them, as well as Windows, and if someone realy wanted, they could find and exploit a hole in Linux. But people seem to forget about that.

Nobody has forgotton about it at all, the fact is it just doesn't happen anywhere near as often on Linux as it does Windows.

Quote:

Now if someone wanted, they might be able to fix the RCP problem and set it up on some website free of charge, but instead these people prefer to use what they find by themselves. Often for personal gain or just to cause haos and to prove that "they can". So, instead of complaining try improving both systems, I'm sure many people will be gratefull.

No, they couldn't. That would require source code access to the RPC portions of Windows. The most they could do is notify Microsoft.

Quote:

May be I don't know as much as some of you guys, but at least I'm trying to look at things from every prospective and don't try to make out that something's perfect when it's not.

It seems more to me like you're using a tunnel-vision approach and assuming a whole lot about how an operating system you've barely used works.

Quote:

Originally posted here by poohsuntzu
No, I honestly don't. Very few people who use Linux, BSD, or Nix complain about window's market share size. OpenSource and Linux isnt here for marketshare. They are here to deliever an operating system that has the capability to be incredibly flexible, moduled, secure, and usable. Marketshare may come, but that's a sideproduct from the origonal goal of Free software and Linux.

To further emphasize the point: Linux and OpenSource in general are NOT directly products of Free Enterprise or Capitalism, they were a product of necessity, and continue to be. This is the fundamental difference between its development and Microsoft's.

Quote:

Originally posted here by poohsuntzu
edit: Granted, when extremely new technology does come out, it will eventually require an upgrade of the kernel because we need to add one new feature for those 20358731486703198 harddrives to use as needed. But by that time, you would want to anyways. Keeping your kernel upgraded, secure, and stable is as essential as keeping Window's core upgraded via Windows Updated.

I beg to differ on this last point. Keeping the kernel updated is absolutely not as important as keeping Windows' core updated via Windows Update if only because the Windows core includes WAY too many things that Linux' kernel does not. In fact, barring a kernel vulnerability and presuming all hardware functions, the only reason to ever upgrade your kernel is speed/stability issues..
This might scare you, but there are plenty of boxes out there running 2.2 still.

Show 40 post(s) from this thread on one page