i dont know if i would do this on amazon, just a compatibility list would do. there are a existing lists out there but find them difficult to use and search. i think if phoronix wanted to do something like that, just start a compatibility list and make it better than whats out there. then users can go find someplace to buy it.
meh. black star green star is kind of stupid to put it lightly. If your someone like me who doesnt care what kind of driver they use as long as it works, then its an insult or in a best case lead buying.
Originally Posted by Qaridarium
Lets be honest. If you cant research your card before you buy it for linux then you probably wont use such a service anyways.
If it's done, it would be tied to a store or stores, for revenue generating reasons.
Originally Posted by mattmatteh
why not honor a good cloused source linux driver?
Originally Posted by L33F3R
but we need to make clear opensource drivers on an long time view is the better way.
in my point of view intel isn't a alternativ becourse high prices and the igp VGAs are slow and pointless.
but intel supports opensource up to 100% LOL!
if opensource means self hurd right intel is the way to go pay highprices for an minimum opensource only support...
in my point of view amd will be much better in the future.
becourse you have the choice open and cloused source driver and specs.
intel has cloused source drivers to but only for macosX/windows
so intel is a joge for linux users.. linux users lose every 3D benchmark against macos only becourse intelshit drivers.
i think if intel brings the same cloused source driver to linux linux will win all benchmarks again macosX..
intel=highprices and a joge only for lose benchmarks its an marketing gag for macos...
i think apple pays intel for this!
sure, lots of people looking for green stars may already have the knowledge, but i still that it would be a good idea.
Originally Posted by L33F3R
... like Music to my ears!
Just want to add my voice saying I like where this thinking is headed!
Please think it out before you get started... because although this ability/service is badly needed it has also been attempted in the past and it seems that each of the projects end up getting mired at similar points in their development.
I'll list out some of the hurdles/issues to consider in designing such a system...
- Do not make a one-off system, instead make a full-fledged open source project. This will help the project spec evolve to cover the changing needs by leveraging the community.
- It would need to be able to take into account the way chipsets change even within individual model numbers.
- It would ideally have a public api so that desktop apps can be written to load up to it. PTS could be _seriously_ leveraged here. Another example would be data collecting web-widgets that could be added to help forums to aggregate user experiences, tips and results.
- It would ideally be federated so that other similar systems can easily and simply share data.
- As difficult as it is, try to be comprehensive... printers, Monitors, mainboards, etc... (This is one area where federation can ease the burden)
- Try to make it an earner for you. I don't think anyone would reasonably mind sponsored tie-ins to services like Amazon (or others) as long as the service doesn't rise or fall on their whim.
- For example, would Best Buy be willing to kick some money your way if people find a device they like and want to know know which Brick'n'Mortars stock it nearby?
The biggest problems I've observed with systems like this is that they end up being isolated, and thus increasingly difficult to maintain.
You're in a great position to fully address this problem and make it a thing of the past.
My final suggestion is to let us help out. I've marveled at how a community so rich in innovation, initiative, and infrastructure has yet to satisfactorily address this problem. Anyway, I'm excited to hear that someone with the resources and vision is taking this on.
Let me know how I can help!
services should be useful first
A quick note...
The discussion about the rating system is something of a distraction. Here's why:
People have different reasons for using linux.
It's just that simple. Ideally a system like this would have all the information necessary to rate among many indexes of interest and concern.
The 'free-ness' of hardware is a high concern for me, but not everyone.
If a school system or retirement home wants to switch to linux but need a binary driver for their printers (for example) I would still be excited for it.
This service needs to have as much correlated open info as possible. So it should be possible to customize to the interests of the person who is searching. I wouldn't expect me and my grandma (over-used cliche) to *want* the same information about what printer to get.
So it's not necessary to ham-fist the interface to impart the importance of the range of software freedoms. In the same vein it would be short-sighted to not direct people to educational resources and explain in simple terms why it matters.
As a community we need to remember that limiting options isn't a substitute for education or comprehension.
I really like the idea..
YACL (Yet another compatibility list) is the worst thing possible. Now I would have to check another site just to see which has the best most up-to-date information. (However it might be good to let compatibility lists pull info from the store)
Echoing some other readers (I'm more comfortable buying hardware from newegg than from amazon)
Ideally we could (eventually) do product comparisons and see real phoronix test suite data on it's performance. Imagine, I'm configuring a system (of all phoronix store approved components) and now I want to see which video card will get me the best performance on ET:QW. If you can integrate that into the buying experience that would rock.
Hopefully I can buy there too (living in Europe).