Linux Needs to Master Hardware to Beat Windows

Page 4 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
@anonymous-kandy-cane: Somebody who isn't intelligent enough to be able to spell correctly probably has a completely worthless opinion about OSes and code. Imagine if you were a developer, you'd never get anything to compile..

naimspace _stuff{

klass _a_klass{
4each(thingie in sumthing)
{
_do_shitt();
}

}
 
hardware is the wrong answer. the answer should have been the total obvious. software. but the real answer is the developers.

there is nothing there in linux for window users to migrate to. a pc gamer wants to play pc games, what is linux doing to say, linux is better then windows in playing pc games. absolutely nothing. so pc gamers don't get linux.

then you have people who don't know about installing a new os. microsoft has a simple wizard installer when installing windows. linux has this command prompt thing, which people don't care or don't know how to use. guess what, people aren't getting linux because linux developers don't care about the needs of these people. so people ditch linux for windows because the installation of windows is easier.

then you have software incompatibilities with linux and windows. well make linux so that it can run those programs. but linux developers don't want to make those programs. so then those people on windows don't migrate to linux.

if linux wants people to get their os, then make is so that it is more like xp. if linux is like xp and can do things like xp, then i will get linux. while linux isn't going to be like xp, then linux will be crap.

the problem with linux is that these people aren't making os for the needs of people. they are making an os for the needs of their own or their own community. if the latter is the case, then it'll stay where it is. if it's the prior, then it'll be a step up. but as of know and forever, it'll be the latter. how do i know it'll be where it is forever, because it has been sometime where linux could have gotten the upper hand but there hasn't been a move to do anything for it.

i don't care for vista or windows 7. here is an opportunity to make an os like windows xp. but there hasn't been.

linux isn't about the customers, it's about the needs of the developers and such. this idea doesn't work because the customers are the ones who bring in the money.
 
people don't really care about windows vista and windows 7. i tried windows 7 and don't like what microsoft did to their operating system.

there is still demand for windows xp because computer sellers like dell are still offering windows xp as an alternative. this tells you that people don't really care about the newer windows vistas and windows 7. this also tells you that it is an opportunity for linux to make a windows xp clone. why not make a linux os program where it'll solve that windows 32/64 bit software problem.
 
[citation][nom]pig_lating_programmer[/nom]@anonymous-kandy-cane: Somebody who isn't intelligent enough to be able to spell correctly probably has a completely worthless opinion about OSes and code.[/citation]


Lol... OSes

 
[citation][nom]pochacco007[/nom]guess what, people aren't getting linux because linux developers don't care about the needs of these people.[/citation]


I think you're being a bit harsh there... It's not that they don't care, it's the engineers have a hard time empathizing with the end users.

Good strong leadership, a consortium of some sorts, that makes standards, checks standards, monitors quality and guides development by recruiting the proper individuals to do the job. Going to be hard to find good quality leadership that will do that for free while they are tied up working 12-18hrs per day at their normal job (like I am).

So you end up with a bunch of talented engineers that either love the challenge, need the ego fueled or simply want to help the community all with very little insight as to what 95% of users need in a computer.
 
Changes in the last few years have been phenomenal and whereas at one point I was just too intimidated to try Linux that feeling is now gone.

My goal is to eventually be using all open source software with Open Office my latest item in the learning curve.

There's still room for improvement where drivers are concerned but to be honest the same can be said about windows.

I mean you have a two year old wireless N adapter and can't find a windows driver for it there's clearly something amiss.

Go Linux!
 
You dont understand the "Real Problem"
The problem is Microsoft Partners preventing Linux users from gaining access to their hardware!
Windows 7 is the greatest example of closed source spyware in the World
The beneficiary of the Windows spyware platform is the Govt.
The Govt will protect its spyware platform at any cost
If Torvold makes a closed source kernel for Linux and enters into a secret agreement with the NSA to hide its code from Linux users the way Microsoft does, then ALL of the hardware out there will suddenly work with Linux!
It is not the resposibility of Linux users to make 3rd party hardware drivers work with Linux if they are intentionally be prevented from doing so by the hardware manufacturers
Get a Clue!
Windows hardware works because the Govt needs it to work for their spyware platform to function properly
 
no... really... the OPERATING system needs to master OPERATING hardware to be successful... thanks for the heads up...

on another note, windows has really upped its game on this front with windows 7 (especially comming from XP) so I think linux has its work cut out for it.

what's open GL support like in linux? can it match DX11 for features and power with available drivers??
 
I find it funny how people say that Linux is not user friendly, as though Linux is a single entity that can have a sweeping judgement passed over it by people who think it doesn't work like Windows and therefore isn't user friendly. If you want to assume Linux is a single entity, then you can only talk about the kernel, not the distributions. I challenge you to compare the user-friendliness of the Linux kernel compared to the Windows NT 6.1 kernel. I have a feeling that neither kernel is going to give you a nice user-friendly, eye-candy-filled GUI to do your work.
 
[citation][nom]Regulas[/nom]That is because MS is buying off game developers to use only their proprietary DirectX. Open GL is still alive and kicking. [/citation]
Linux fanboy much? The reason OpenGl is rarely used today is because the API is to my understanding cluttered, inefficient, and lacks most of Directx's features.
 
[citation][nom]pcxt21[/nom]Linux fanboy much? The reason OpenGl is rarely used today is because the API is to my understanding cluttered, inefficient, and lacks most of Directx's features.[/citation]
OpenGL is stuck in a viscous cycle. It won't advance unless it sees greater uptake, but it won't get greater uptake until it can compete with the features offered by DirectX.
 
Linux Vs windows..
Another funny debate.

Solution:
Produce the latest hardware with linux drivers. (that will make a huge change to the computing industry).. Instantly.

Also . a further note.
The linux community suffers from OVER Software. Yes you read it right.
Linux has way to much software. The windows community has it's head in the sand on this. . So one of the bigger problems for Linux is simplifying or filtering out the tons and tons of software it has.
Realy.. this is not a good thing. rather bad.

 
[citation][nom]pig_lating_programmer[/nom]@anonymous-kandy-cane: Somebody who isn't intelligent enough to be able to spell correctly probably has a completely worthless opinion about OSes and code. Imagine if you were a developer, you'd never get anything to compile..naimspace _stuff{klass _a_klass{4each(thingie in sumthing){_do_shitt();}}[/citation]

you could if you made some defines. don't need to be able to spell correctly to program, just need to be able to spell consistently
 
You must treat each major distro of Linux as a separate OS as they often have their own standards for configuration files and how they treat startup scripts / services. That fact alone means developers need to sink massive money / time (which is just another form of money) into developing / testing for each major distro. And with no single player at the helm there is no one for a software developer to "partner" with to get compatibility verification. The most they can say is "with our installation, on our hardware, with out operating environment, on our kernel, with our specific configuration, it should work". That's too much to ask of a company for a relatively small user base.

As for OpenGL, the problem is that its "stuck in committee". There is no single person / group of people developing it. Its actually a group of software / HW manufactures, a majority of which actually work with CAD/CAM and not the gaming works. Those guys do NOT want to have to reprogram their entire code base to support gaming features, they don't want to bog down the protocol for things like shaders. For them the protocol is about being fast and efficient for designing and manufacturing in robotic factories and such. So OpenGL will never have the features of DirectX. It was to rely on extensions provided by GFX card manufactures to have advanced features, but then these tend to be GFX card specific (NVidia tends to be stronger about OpenGL support over ATI).

I personally feel OpenGL is a better way to interface with GFX cards, but until someone takes it and develops their own system that isn't associated with the group it won't go anywhere.

And for that same reason, until someone in the Linux community develops a single platform / standard for basically "everything" it won't ever be mainstream. RHEL is the standard for most corporate linux systems (their usually compatible with RHEL scripts / configurations), need something similar for the desktop space.
 
forget mastering hardware first. they will not get anywhere until they make it more user friendly and easy to use like windows so the majority of PC users will see it as a alternative that is the same as windows.

there is no way they can master hardware because there are a million distros of it out there and many more from the script kiddies
 
I only deliver Ubuntu systems based on Intel/AMD/Nvidia chip sets. Never had a hickupm with this hardware configuration. Always works like a dream even when changing motherboards. Linux is just there. I don't know about the windoze fraternity, as far as i am concerned they just are just ripping users off and refuse buying anything with a windoze label.
 
I have an idea about improving Linux (and OSes in general) that I would like Mr. Curtis to comment on directly. A lot of posts here have already commented on the problems that my idea seeks to solve, so I wont go too far into why I think it's a central problem. I think the installer and build pipeline problems are just as important as (and directly related to) driver support issues.

I think Linux needs a "Universal Installer" format, similar to the "Universal Installer" that Apple used to help transition Mac OSX from PowerPC to x86 architectures, and later from 32-bit x86 to 64-bit. Linux not only has hardware CPU platform diversity to deal with, just like OSX did, but also kernel detail diversity. If you hear me out, I think the particular "Universal Installer" I have in mind would help with the hardware driver support issues you speak of as well.

I don't think current package management systems are sufficient, in part because they're solving the wrong problem. They are still solving (and sometimes failing) at the "library hell" problem, where conflicting library versions cause unforeseen errors in the software that shares those libraries. Windows has multi-part solutions to their "DLL Hell" counterpart, which are still sometimes insufficient, so many Windows installers end up just putting many versions of the same driver in their own respective Programs folders.
I think a better solution is to have a version controlled shared library database. Instead of forcing all local software to upgrade a shared library when one item does, you keep a "delta" or "snapshot" of the older version(s). Software that doesn't yet support the newest version of the library then just keeps a pointer to the old version, that can be re-pointed to the new version after the next upgrade, or compatibility confirmation message. Older libraries with 0 remaining pointers can be deleted in a "deprecated collection" pass later. Disk space usage can also be minimized by storing old library versions as deltas from the newest matching version. A file system that supports snapshots and symlinks would be sufficient for this solution, and some simple library install and management tools could handle pointer generation and cleanup tasks.

The other problem with any installation, that I personally think is even more important than "library hell", is kernel+CPU diversity. I don't think it's appropriate to distribute kernel+CPU architecture specific binaries, and expect them to "just work," anymore. Where I have high hopes of this changing is via compiler toolchain projects like LLVM. LLVM provides a platform agnostic bytecode format and VM that can be targeted by any language compiler, including for proprietary code. That bytecode format can then be sent to any other platform, and the piece of LLVM that compiles bytecode into local/native machine code can complete that step at install time. Proprietary driver vendors, who sometimes don't have full control over their drivers due to 3rd party license arrangements (nVidia comes to mind), can send proprietary bits as obfuscated bytecode, and open-source-ize the rest. They can also keep their own code proprietary via obfuscating the bytecode, so that OSS community support becomes a separate issue that doesn't hamper timely distribution. This also gives major proprietary software vendors (like Adobe) a much easier distribution method for targeting all Linux releases.
I can foresee some issues that might require install scripting with the use of platform-specific environment variables, but that isn't anything new to installers, such as the Windows MSI installer tool. I see the resulting "Universal Installer" binary as a zipped package of LLVM bytecode files, libary dependencies, and wizards/scripts that define operation order and output paths, using local environment variables and platform-specific conventions. Those scripts could even utilize local environment knowledge to tweak compiler output using platform-specific optimizations and methods.

The same LLVM "Universal Installer" toolchain could even be created/ported for other OSes, like Windows and OSX. Then a developer can use the LLVM tool chain to finally live the dream of all coders: write once, distribute everywhere, and run anywhere.
 
Packagers: there are currently two package formats: RPM (Red Hat/Fedora, Suse, Mandriva...) and DEB (Debian, Ubuntu...). Interestigly, there is already software out there that can install DEB packages on a RPM machine, and vice versa. So, the 'packaging format' is a false problem (and please note that this problem also exists in Windows: you need to ship a 3 Mb binary with your installer to be sure that the system you want to install in can handle the specific version of MSI you're targeting it with).

Moreover, for anybody who has ever installed software in Windows that required this or that toolkit, Linux packages have one definite advantages. Let's take an example.

You have this app, the only one on your system that need .Net at all, and this one requires .Net 3.5. It also is the only one to require MSVC++ 2009. Now, you're done with it, and you want to remove it completely, as the 400+ Mb used up by all those toolkits take up too much room for comfort.

You have to manually remove .Net 3.5, then 3.0, then 2.0; then, remove MSVC++. You then need a registry cleaner (because a bunch of unused keys stayed behind) and defragmenter (because those keys were numerous) before you can finally remove temporary files left behind by the installer. That is, if everything goes according to plan - otherwise you'll need a brute force uninstaller.

On the other hand, on Linux, if you remove one package, all its dependencies will be removed too; more recent ones even allow you to clean up installed toolkits that are no longer in use. You can also use a well made Fedora package on Mandriva or Suse, provided the packager didn't forget to include file requirements - so that the package manager can find out the closest equivalents.

About the command line: as far as I've seen, on recent distros, double-clicking on a downloaded archive will prompt you to install it, resolve dependencies, and then complete successfully or tell you what files are missing to conclude the install.
 
[citation][nom]ronch79[/nom]No, not just hardware drivers, but the lack of software is the primary reason why most people stay away from Linux. [/citation]
Yet they don't complain about the lack of software on Windows? There's far more for Linux, but most people will never have a use for it all, as they don't with Windows software.

[citation][nom]ronch79[/nom]Their complaint is that the software they're running can't run under Linux.[/citation]
Well of course it can't. Windows software should be run on Windows. If you want to run Linux, use Linux software. There's no point switching OS to one with a totally different philosophy if you don't want to change your way of doing things. If a person is stuck in their ways, they should stay with Windows. It's a much better option than whining about Linux's "problems."

[citation][nom]ronch79[/nom]As for looking for a YM equivalent for Linux, most users simply wouldn't bother.[/citation]
There's one pre-installed with most distributions, which handles almost every major messenger service (assuming, by YM, you meant Yahoo! Messenger).
 
I have had no problems running any Linux system on Dell computers including the Vostro that I am typing from. I actually disagree that it is the hardware, I believe that it is software that holds people back from using Linux systems. Because most software has been on Windows for so many years it is mature. From Audio recording, gaming, to Office software just to name a few. 9/10 times someone stays with Windows because they can't find a suitable Linux replacement program. This is also why people are stuck dual booting, setting up virtual machines, or running programs through wine. It is all about the software if Linux wants to get real.
 
[citation][nom]kyeana[/nom]What we really need is ATI linux drivers. If we could get that i would be very content.[/citation]

Ummm... I just installed the newest Mint. ATI drivers out of the box, supporting HDMI audio. What more could you ask for?


People who are to stupid to learn the terminal commands, have no business using Linux. They can pay to use M$ and be happy with it.

Whats sad is that I took me about a day to get comfortable with it. Seriously.... Not THAT hard.

So people can whine and cry all they want. Go get their refund for their Linux purchase (lol), and complain some more about how its not Windows.

I wanted something free, stable, secure, customizable, and has great community support.

I went with Linux. Hard decision! I know...

Background:
I've been using windows since the dawn of it. I just recently installed Linux Mint 8 on my computer and only boot windows when I want to play a game. It was fun to install, and get to know. I opened my mind, accepted that it is NOT WINDOWS, and used it for what it is. A solution for a need that everyone can contribute to. Thats what Linux is all about.

So far, I'm happy. It does everything I want it to do, and I don't have to worry about the Bull Sh* that M$ dishes out to us all the time.
 
I agree about linux being useful if it supported the majority of hardware on the market. It is the perfect solution for many uses such as having a data server on the home network or even for HTPC use. I was thinking about using for HTPC duty, but there were no drivers for my TV tuner card so no dice. I like to build my own systems so I have to pay a heavy microsoft tax, a free alternative would be great. I just need the free alternative to meet my needs which at this point it does not. Chicken and the egg I think, nobody uses it because it isn't well supported and it isnt well supported because nobody uses it. :) Of course I think I will always need windows on my desktop for gaming duty.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom