[minthaka]

I've got an idea: what about opening a hardware database - tab in Linuxforums? I'll try to explain what I mean:
1. All of us have encountered some troubles installing hardware parts like printers, scanners, etc.
2. I would like to approach differently. I can offer the method I used to successfully install my hardware parts. So I can give step-by-step instructions, how to make parts work. I can describe it on distro's I've used.
3. If we'd all do the same, we'd create a huge database about wich hardware to buy. For example: somebody would search about Samsung printers. He'd type in Samsung, and would get back as an answer a list of supported types and also of the list of distributions supporting that particular series. He'd also have a clearly detailed step-by-step description of how to install on a particular distribution, where to get the driver from. I'm willing to contribute with my experience, if others would join.

I'm not as skilled to design the page within Linuxforums domain, but I gave my idea. I think it would help invite people into Linux-world.

[ozar]

Hi, minthaka

Any such database would have to be approved and then implemented by Core Wizard, as he's the only one with the permissions required to implement it. I'm sure he'll give your suggestion consideration as soon as he sees your thread.

Thanks for the feedback.

oz

[Freston]

I think that is a great idea!

It is, however, tried and failed IMHO... not on this board, but on several websites.

Failed, I say? That is not because of lack of effort. Why do these attempts fail then?
Well, I must add, they fail in my eyes. My taste, so to speak. The thing is, there are 1,172 printers supported by hplip. But not all of them, not even all HP printers (as I understand).

Also, some hardware is partially supported. Take my PB EasyNote 1904 (bad choice). It works, but not without appending <acpi=force pci=noacpi> to the kernel. Which makes my lappy can't switch on the printer, I have to do that manually. Or is that the driver?

The complicated thing here, as you see, is that either you don't get these questions answered, or even worst! these answers clutter the info to such an extend that you can't see the forest through the trees.


There are so many models, so many, and revisions on models. A typical line of notebooks looks like this:
Cheap-o-line 4450
Cheap-o-line 4460
Cheap-o-line 4470
Cheap-o-line 4475
Cheap-o-line 4475-c
Cheap-o-line 4475-r
Cheap-o-line 4490
Budgetmodel 23991
Budgetmodel 23992
Budgetmodel 24235
Budgetmodel 223-33/DDX
Budgetmodel 228-43-k-32
SuperAverage DeLuxe
SuperAverage Uber
SuperAverage Special
SuperAverage With a Twist
SuperAverage 4D
Uber-great Gaming
Uber-great Multi-media
Uber-great Office
Uber-great Home edition


And that is only one manufacturer. Do you know if the TV-tuner in the 223-33/DDX works under Linux? Is it the same TV-tuner as in the Multi-media model? And if you don't know, the guy at the electro store sure as Bob doesn't either. Not to mention that at every store they have different models of the same manufacturer. So if you decided at home googeling for hours on which one you want, they nowhere have that model but ample of models like it.

This, my friend, is a challenge to be overcome. I am more than willing to think this through. My hardware knowledge may be below par on this here board... but I may be able to look at this from another point of view: 'presentation of data'. The complexity of that subject is often underestimated.

Actually, thinking about it, that could be fun... in a geekish sort of way


Yeah. Maybe we could sketch out some of the rough outlines. I have to work this weekend though. But I'll give it some thought.

[Core Wizard]

I like this idea. I will give it some serious thoughts after some major changes take place on these forums, that would make them much more user friendly. This forum is in for a redesign within the next month, as well as a few new sections, wait and see . However, if you can sketch up exactly how you see this, I'd be more than glad to hear it...

Mark.

[minthaka]

Let's see:
I have a Mustek 1200UB-Plus scanner that, along with the installation's procedure I would like to add to the database.
1. I would click some button "Add hardware description".
2. I could choose "Category" (Printers, Scanners, Cameras,....).
3. After that I would have a field (for a new one) or a combo-box (for the existing ones) with "Manufacturers".
4. Next would be a a field/combo-box with new/existing models.
5. I would get a form to fill with distribution's name, and a field to add step-by-step instruction for installation. I could alternatively do the same thing for different distros, -> button "New distro".
There would be maybe even a warning that I'm attempting to add a description which already exist for that Category/Manufacturer/Model/Distribution.
6. Having finished the typing I would "Save" the data, and it would be available for others to search for.

A similar procedure would be the search :
Select Category:
Select Manufacturer:
Select Model (exact) :
There would be listed (if it's supported) on which distro is supported, how to install, and where to get driver/firmware from.

That's it!
We would only add our knowledge, or experience! There's no need to know everything, every subtype or manufacturer, just share what you know, and let others use your experience.
I think it would be a very powerful tool.
So if I saw a scanner in the shop, I could check whether to buy it or not. I didn't have to search over hours on the internet. I would have it ALL in one place + with possibilities to contact the person gave the instruction.

Do you like it? Maybe some manufacturers would feel themselves pleased to be find in such a frequented place!

[Freston]

Yeah I like it . That is exactly how it should work, from an interface point of view. But, I also think the matter is a bit more complex than how you state it. And I'll tell you why. This is done over and over. But what you'll end up with is a tuned down version of this.

And this is what I mean with not being able to see the forest through the trees. Don't get me wrong, I respect the huge effort that went into such a project. And all those make/model specific howto's can be a great help for all the people with the same make/model. So the effort isn't wasted, oh no! Far from it. But it's a help for someone who already has the make/model listed. It's not really helping someone deciding upon which make/model to buy. In the fast pace of computer technology, a setup like that is always behind on the facts.


--- Consider this ---

If you'll come with me to the corner electro store. This is where most people buy their hardware. As you can see, every electro store has the same brands, only slightly different models. What do we have?
Of course, two or three HP's to choose from. Also two or three Acers, three PB's, a Compac or two. Probably some home brand (cheap). And the mandatory Toshiba.

Now how do I know which one of these will work? I have a fair chance that all of them will work just fine. But, if I'll find out after installing Linux, that the Wireless/TV-tuner/Graphics/Whatever doesn't work, I cannot return the machine. I'm really disappointed and out of a lot of money. So I don't depend on my fair chance, instead I proceed with caution.

---/Consider this---

This situation is faced daily by a lot of people. And, as we are *nix people, we probably think these issues should not be faced by humans, but rather by an online engine of some sort.






I'm at high risk of exposing my limited knowledge on the subject here, but I think a database should be able to break down a make/model into it's components and view their 'support-status'.

A database (I'm typing faster as I'm pressed for time) can store an almost infinitely large number of components. When a new make/model gets feeded into the database, the database can reference this make/models components against what is already in it. If for example all the components have been proved to work, there is no need to actually test this new make/model.

For example: if we all know the new Toplap UberGamer 3000XX has the NviaVi 384993 card and the LinkRa 9090 wireless card, and we all know these components are well supported, then it's safe to say the new Toplap UberGamer 3000XX is supported.

A database like that can also deal with questions like: 'are all DVD-RW's supported?', for example, I have come to understand that each and every Bluetooth device works under Linux.

----

So, if you're still with me... I have two questions. First, is breaking down a make/model into it's components a safe approach to estimate the workability of a particular model under Linux?
And second, how will all this wisdom get feeded into a database?
(At first human effort I'm afraid, although with the manufacturers support this can be 99,9% automated in the end) Then we'll need a way to tell the manufacturer something to the extend of "If you want your new Ubergamer 3000XX in our database, you'll have to email us the specs in this-and-this format. Then the computer can handle the rest. And we have ourselves some innovation.

*Looks at the clock* Think about that. I'm off to work Feedback and suggestions will receive a warm welcome

EDIT: Great initiative minthaka!