How to help / support Linux ? - your idea's

[crazycheese]

One thing would be for users not to bail out as soon as a piece of hardware/software isn't supported in Linux.

Find other people having the same issue and collectively contact the manufacturer.

How about you purchasing hardware and manufacturer simply refusing or puting you on the long todo list, like in case of amd?
I think the thing is dead simple - buy only working linux hardware - hardware from manufacturer that does not hestate to speak to linux user or invest their money in their development.

[saysilence]

Nice idea but the problem with that it doesn't help get projects established. It is a "the rich get richer and the poorer stay poor" scheme. Canada imposed a levy like that on CD media that was supposed to help out music artists. Unfortunately the only ones that see any kind of dividend is the already well established artist while the up and comers and smaller artists get nothing.

Well i have not explain the whole idea, the foundation may be used to fund project that needs to be done and by setting some upper and lower limits of that revenue, such a foundation can make sure that everyone gets something.

About the "the rich get richer and the poorer stay poor" i do prefer to have less but good software instead of many yet worthless ones. And it usually happen that good software quickly gets popular, even if it's new.

[saysilence]

This wont work.
- Money is not put on specific target, but is granted to some organisation "in advance".
- License overhead and its essence crap
- Polarization around one entity instead of targeted spread
- Many many people do not have 1$, do not wish to spend $1 for sake of "support" instead of return or need $1 to pay food. Rich will laugh, able will question the reasoning, poor will not be able to manage it.

People are really really stupid beings that prefer to pay reasonable price for the job be done and they not going with their head into that; and they love colorful stupid slogans (Lookup ubuntu marketshare - they are barely inventing anything, but "talk talk talk").

Sorry, my opinion. Not that this is to stop you. But you might one day recieve a hatemail from RMS.

- Money is not put on specific target, but is granted to some organisation "in advance".

To put money on some specific target you first have to have money. Granting a small sum to an organization which can be trusted that will gather a larger sum ( lets say 50 000 000 users pay one buck ) you get a capital that can be redistributed to the important projects that needs support.

- License overhead and its essence crap

A well constructed license is essential to protect the intellectual property of the developers. A port of the gathered money could by used by the foundation to prosecute those who breach the license and protect you (the developer) from let's say, someone re-branding your code and selling it as their own.

- Polarization around one entity instead of targeted spread

Please study the mining of "Synergy" and "Unification". Gathering the whole potential of community in one place can create a leverage able to push development forward.

- Many many people do not have 1$, do not wish to spend $1 for sake of "support" instead of return or need $1 to pay food. Rich will laugh, able will question the reasoning, poor will not be able to manage it.


Have you ever heard about Humble Indy bundle? Many people are not like you, many would willingly pay one dollar and much more to help the community. And people that can't pay one dollar a year usually don't have computers and they usually don't care for open software.

[acolomb]

How about you purchasing hardware and manufacturer simply refusing or puting you on the long todo list, like in case of amd?
I think the thing is dead simple - buy only working linux hardware - hardware from manufacturer that does not hestate to speak to linux user or invest their money in their development.

That scheme relies on the influence of the masses. Unfortunately, the group of linux users is just a fraction of the potential hardware buyers, which was the problem in the first place. From my experience, the desired market pressure is purely wishful thinking. No change in hardware vendors' strategy based on it.

[dacha]

1. Hardware test farm

Hardware support problems, and regressions, are numerous. The current system is to release a new version of (say) Ubuntu with minimal testing, then rely on bug reports from users to find and fix problems. In other words, only users' bad experiences *sometimes* cause progress, and nothing stops a regression.

Instead, I suggest the construction of a open-source "test farm". This test farm has tens of thousands of machines, with the widest range of hardware under the sun. Every day or so, the kernel head is deployed to all the machines in the test farm, and a thorough battery of tests is run. These tests including everything, kernel bootability, OpenGL frame rate, network/modem connectivity, mouse/keyboard input, disk I/O, CPU performance benchmarks, errors in dmesg, system stability, etc.

The benefits are significant. A daily report with any problems/regressions can be automatically posted to lkml so they can fix the regressions. The reports also give a good indication of the stability of the kernel, so distros can pick a stable kernel version to release.

The test farm doesn't have to cost much. Volunteers could connect their own machines to some central server running the software. Testing old systems is valuable too.

2. Unsupported hardware count

Distros should use a combination of "lspci -n", "lsusb -v", and user bug reports, to count how many users are running an unsupported piece of hardware, then report that to lkml and the hardware manufacturer.

What sounds more convincing, one user saying "please write a linux driver", or Canonical saying "250000 users run your hardware on Linux, you are getting bad PR with your current Linux driver situation"?

3. "Supported by Linux" label program

For when you buy hardware.

4. Linux app store

Need I say more?

5. Device integration

Get the latest gadgets (iPod, iPhone, Android Phones, Blackberries) working perfectly on Linux.

[crazycheese]

That scheme relies on the influence of the masses. Unfortunately, the group of linux users is just a fraction of the potential hardware buyers, which was the problem in the first place. From my experience, the desired market pressure is purely wishful thinking. No change in hardware vendors' strategy based on it.

This scheme is already working. Some companies are less ignorant than the others.

[crazycheese]

To put money on some specific target you first have to have money. Granting a small sum to an organization which can be trusted that will gather a larger sum ( lets say 50 000 000 users pay one buck ) you get a capital that can be redistributed to the important projects that needs support.

For what reason? Why not to allow people vote with money themself on clearly defined targets?

A well constructed license is essential to protect the intellectual property of the developers. A port of the gathered money could by used by the foundation to prosecute those who breach the license and protect you (the developer) from let's say, someone re-branding your code and selling it as their own.

You already have "proprietary" for say, since AT&T unix deal. You already have proprietary windows. What are you trying to protect on linux? Something that EFF already deals with?

Please study the mining of "Synergy" and "Unification". Gathering the whole potential of community in one place can create a leverage able to push development forward.

Like I sayd, you are reinventing microsoft.

Have you ever heard about Humble Indy bundle? Many people are not like you, many would willingly pay one dollar and much more to help the community. And people that can't pay one dollar a year usually don't have computers and they usually don't care for open software.

Of course I have heard. It is an attempt to
- sell more proprietary code
- attract more proprietary companies to linux, as additional market

It has nothing to do with opensource, nothing to do with opensource development and it is not long-term.

[dacha]

That scheme relies on the influence of the masses. Unfortunately, the group of linux users is just a fraction of the potential hardware buyers, which was the problem in the first place. From my experience, the desired market pressure is purely wishful thinking. No change in hardware vendors' strategy based on it.

How about:

1. Hardware test farm

Construct a farm of thousands of machines with a wide variety of hardware, and every day, build the kernel head and run a battery of extensive tests (kernel bootability, OpenGL frame rate, network connectivity, CPU benchmarks, disk I/O, USB I/O, suspend/resume, dmesg errors, system stability) and mail the report to lkml.

This simple step would create a major improvement in kernel quality. Hardware regressions would be obvious within 24 hours. Distributions would only release kernel versions that work with all hardware. Developers wouldn't have to guess how their driver patch affects other variations of the device that are in the field.

The test farm doesn't have to cost a lot. Volunteers could run an application on their computer that connects to a central server and downloads and runs the necessary tests.

2. Unsupported device count

Distros should automatically scan your PC for PCI and USB devices that aren't supported and then report these to the manufacturer.

Canonical saying "250000 users are costing you bad PR because they can't use your device on Linux" seems a lot more convincing than a single user's "Please make me a Linux driver".

3. "Works on Linux" labels on hardware

4. Linux app store

5. Get Apple's gadgets, Blackberries and Android phones working perfectly on Linux

[masinick]

I believe in choices and alternatives in freely available software

I believe in choices and alternatives in freely available software, and I believe those choices should apply to which distributions we use, and how we contribute - and even if we choose to contribute or not. Personally, I see it advantageous to contribute in one or more ways to free software projects, but that, by its very nature, is a choice.

I believe in the choice of whether or not to financially contribute to projects, whether or not to test and provide feedback to projects, whether or not to suggest new features, removal of aging features, coding, testing, helping, documenting, managing projects, and so forth.

Two of the easiest ways a relative newcomer can get involved is to try out new systems, and report back their findings. In order for that to be helpful, it takes a little - but not an overwhelming - amount of detail and discipline. Simply, the best thing to do is to write down what you did, the steps and order in which you did them, the outcome you encountered, and the reasons why you feel that outcome is either good - and praise is always nice to see for a job well done, or constructive criticism, when you believe that the expected results ought to be different, and the reasons why you'd like to see them different.

There are plenty of people who write to forums. Unfortunately, only some of them follow any kind of discipline when writing in. Statements about likes, dislikes, expectations, and so forth, without any kind of substantiation tend to be time wasters, rather than helpful comments. Opinions certainly have value, but to me, the value is exponentially increased when the reasons behind the opinions - and the value the opinions hold. make them much more useful.

For example, if you find that you download a particular distribution and it will not boot, a good problem report might include some or all of the following information: 1. Where did you download the image from (which site or mirror location)? 2. Did you perform any kind of checksum or other form of media verification, and did it match any published information? 3. How did you create the boot media - (what speed was the image creation, if using a CD or DVD burning application, for instance)? 4. What are the hardware specifications of the system or systems you attempted to use? Include graphics driver information, disk type and capacity, model numbers, and any other identifying information that could assist in locating and identifying possible hardware compatibility issues. 5. Describe the process you used to install or test the software in question. Include the steps you performed and the results you encountered. If there are unexpected results, briefly describe what you would expect to see in order to be completely satisfied.

These are some ideas; there could be other statements, other ways to be helpful, but these are some of the most helpful things that most of us ought to be able to do that could considerably help others - those who answer questions, as well as those who have encountered similar issues. Modeling our reports in helpful ways can be useful to others who are learning.

[PsynoKhi0]

well thats like saying someone is in the middle of a gun fight and just decides to run away because they don't feel like being killed - not everything is the right choice for everyone. linux isn't the best os for the average person and if they're caught in an unexpected mess then they're not going to waste time trying to fix something they know nothing about.

Not every piece of software is a matter or life and death though
Leaving aside the "best os for the average person" argument, if, say, your business relies on a computer-related tech that isn't available on anything else than a particular OS, you may have other priorities than expecting the manufacturer to support your OS of choice.
What I mean is that letting your voice heard so that suppliers know you want a Linux port is still some kind of support, one that any user can provide.

manufacturers don't usually make the drivers for linux, so they don't have a way of helping anyone. video cards and some wireless cards are really the only common drivers made by the manufacturers. you idea isn't a bad one but remember that one of the greatest weaknesses of linux is a lack of professional and/or dedicated support.

They can release specs.
There might not be any centralized support for the GNU/Linux ecosystem as a whole, there is a dedicated community instead
Just curious, how many would apply if a Big Name OEM among e.g. PC manufacturers went out and said: "Look, we've got that new model coming out soon, we'd like to offer Linux as an alternative on the base install, but we can't test each an every distribution out there, so we'd ask users to test run the hardware for us, under NDA"?
I'd sure raise my hand at once