H-node
- Login o registrati per inviare commenti
Hi everyone.
I just wanted to ask you whether you have checked your hardware's compatibility with free software in h-node.org? If you have not, I can recommend you to do so. H-node keeps track of any kind of hardware that is compatible with free software only. Its easy to contribute back. Just go to their Search page and input your lspci -vmmnn output there.
It would be awesome if you could contribute.
I added my Netbook (Asus 1001PX), desktop onboard graphics (VIA Unichrome Pro) and TV card (LifeView FlyVideo 2000) and a few webcameras and wireless cards to h-node. I didn't manage to add my previous laptop which died though (although I did add its webcamera and graphics card) :(
Please report also hardware that doesn't work, often it's similarly named to working one.
There is h-client, it makes reporting non-laptop devices easier.
I added my puter and my gf's too!
So far, I added a Brother mfcj220 printer, and a pvr-150 capture card. I agree with lammie87. H-node has the potential to become an outstanding source of information concerning what works and doesn't work with free software. In my experience, H-node made purchasing decisions worthwhile and spot on.
I like the idea of h-node. What I don't like is how people are using it as the be all for purchasing hardware. Just because something works doesn't mean it works well or has chipset/manufacturer that supports it (and especially properly).
For example it makes a lot more sense to go with an Atheros wireless PCIe chipset over the equivalent Broadcom chipset.
While both are freedom friendly Broadcom has not put forth the same level of support for free software. Until recently Braodcom didn't even offer a driver, documentation, or any cooperation at all on GNU/Linux.
The same is true for many other types of product. HP for instance does an excelent job providing documentation. While this documentation is difficult to read for most GNU/Linux users it's at least available. It can be utilized to make educated decisions about which printers to buy. They go beyond just saying it works/doesn't work though. They write, release, and support the driver. In comparison most other manufacturers end up with GNU/Linux support purely because of standards compliance on higher end models. In many/most instances the lower end models from other manufacturers (those without standards compliance) are left with no driver at all for GNU/Linux or a poorly written non-free driver. This driver generally doesn't work across distributions or versions.
These are just two examples. There are many others (NVIDIA, etc).
I think the problem is that the actual use of this information and the goal/point of the project are not the same.
Clear information on the issues with various companies/chipsets/etc I think would help fix this problem. A lot of this information can be generalized enough for people to make educated decisions. At least provided that they are technically savvy to follow it.
I agree this is better buying harware from manufacturers that are the best free software "friends". But how can you find this information? I personally don't know how. But if you know, you might be able to create and maintain a list of manufacturers and give them a notation about their friendliness to free software. And this, in conjunction with h-node, would provide good informations to choose free software-compliant hardware, don't you think?
That is because there are few people who deal with these issues on a day to day basis and are supporting a large user base. If I hadn't worked for a distribution in 2005 with an opposing perspective to free software I wouldn't have realized just how big a problem it was. When I founded ThinkPenguin I had to make a decision. That decision was to support non-free software or not. It was a technical decision at the core. Not an ethical one. This is not to say that was the only reason. It was never a purely technical decision. It was also an ethical decision. The first distribution we attempted to support was GnewSense. After which was Trisquel. All this despite there being no real demand.
To continue on. The company I worked for tried to and failed miserably at supporting non-free software. Unfortunately Ubuntu and others are making the same mistakes. The difference between Canonical and the distributor I worked for was one of public relations. In fact many of the developers from two different companies during that time period whom focused on non-free components or in cohesion with non-free software ended up at Canonical. One of the developers I worked with was obsessed with Apple (as can be seen in Ubuntu) and explicitly stated GNU/Linux didn't matter. This was said to me personally. I wouldn't likely have realized the extent of the non-free software problem either had I not seen first hand just how badly it was hurting GNU/Linux adoption. His actions and influence are still being felt. However he is certainly NOT the only person to blame. He doesn't work in a bubble. Mark Shuttleworth, Kevin Carmony, Andreas Typaldos, Linus Torvalds, and many others are part of the problems facing users.
We don't need non-free software or a stable ABI. We should make some changes in order to better support GNU/Linux and free software users though. The current setup works OK provided we have a company like ThinkPenguin although it could work better I believe with more cooperation between various companies and the free software community. That is chipset companies like Atheros, hardware manufactures, distributors like Trisquel, Ubuntu, Linux Mint, and retailers.
Here is an example of why we need to do things differently. It's also why h-node is ultimately not the way to solve the problem even with more documentation on how to choose your hardware.
How to pick a printer:
HP is the only printer manufacturer which is currently providing comprehensive documentation on printers and support for GNU/Linux AND free operating systems. This documentation is very cryptic though. However it's sufficient to figuring out which models to look at if you know how to interpret it.
I'm going to start by visiting the HP web site. There are lots of current models listed. These printers are readily available from thousands of retail stores and online shops around the world.
Once you find the printer that you are thinking about buying you need to check the HP's recommended printer list found here:
http://hplipopensource.com/hplip-web/recommended.html
If you click on a particular model there is lots of information to check out about driver compatibility, distribution compatibility, and more. This is NOT a good way to determine if your printer a particular printer is well supported. HP's recommendations are not very good. In fact this information is even misleading to those who don't understand it.
On this page there lists the HPLIP version. This is a critical component to determining which distributions and versions have out of the box support. The distributions listed on this page do not all work out of the box. This is only the versions which are compatible with the HPLIP driver. However this is not the version which ships with the distribution. It only means that a version of HPLIP driver can be installed. That procedure is complicated and I would not recommend it. While it may be faster to install this driver on GNU/Linux than it would the equivalent Microsoft Windows driver it requires more technical knowledge than most users have.
So.. to continue. The first thing you want to check is that there is no “Driver plug-in” required and that the unit does not require a “firmware download”. This information is found at the bottom. If this is the case the next thing one would need to check is that the support level is “Full” and it's a “Recommended” model. This information is at the very top.
Now that you know this you need to find out if the printer is supported by the version of the distribution you are using and not just by the driver HP makes available. To do that you need to find the “Minimum HPLIP version”. This is also at the very top of the page.
Now open up a new tab in your web browser and visit http://packages.yourdistribution.com/ or search through your package management software for the version of HPLIP that ships with your distribution. packages.yourdistribution.com only works for debian derived distributions. There are equivalents for at least some other distributions.
If using the web site scroll down to “Search package directories” and enter HPLIP. Select the “Distribution” version you have installed. If you don't know what version you have you can find it under System Settings → Details (this may differ from one distribution or version to the next). Now you will need to translate the number into a name.
To translate it open up a new tab in your web browser. Search for wikipedia + distribution name. Select the resulting link pointing to the wikipedia distributions information page. Here you can generally find a section titled Releases. It'll tell you the code name for the version number. Find it and go back to the packages.yourdistribution.com page.
Now you can select the distribution version based on the code name for your distribution. Hit search. This will result in the listing of packages with hplip in the title. The first result is usually just “hplip”. This is the one we are looking for. Under it you will see the hplip version number. This must be greater than the version from the hplipopensource.com page. Otherwise it won't have out of the box support.
Now here is where the problem is introduced and one of the reasons why libre.thinkpenguin.com exists. The currently available printers are not necessarily the ones supported by the long term support releases that are recommended for most users. For that you need a company specializing in supporting GNU/Linux and free software. Once you realize this and choose the right printer you'll have a better experience on GNU/Linux than with any other operating system. There are no drivers to install and devices will work out of the box. Simply plug the printer in and hit print.
One last thing. There are multiple versions of these printers. Not all are supported by the same HPLIP version. If you order one particular model you are likely to end up with a printer that does not work with the version of the distribution you are using.
This is another reason h-node doesn't work. The model numbers do not equate to the chipsets or variations in the hardware which is sold. The problem is also prevent in USB wifi adapters and other types of hardware.
I could write a book on how to pick out hardware. These general guidelines though will at least point you in the right direction. There are still yet other issues from quality control to availability that are not solved by knowing the right hardware to get.
Getting the right hardware does happen. Frequently almost by chance. There have been a lot of errors that can't be corrected due to the complexity of such databases and the lack of knowledge from ordinary users. Your own experience with h-node and success at utilizing it is not applicable to the larger segment.
I've seen in action numerous database projects like h-node fail. They work at first and die out of usefulness quickly thereafter. We have many many many utterly useless databases from various distributors over the years.
And again- this information on h-node is useful to persons such as myself who can utilize it effectively (although not successfully every time due to these and described limitations amongst others). It's not something I'd recommend to your average user or even most technical users.
That was great info, thanks a lot. Just used it to buy a laser printer.
There is an even simpler way to determine your hplip version.
$ dpkg -l hplip
Especially since the yourdistribution.com
URL doesn't seem to exist any more.
I'm pretty sure he meant e.g. http://packages.trisquel.info.
This is my addition:
http://h-node.org/notebooks/view/en/524/Alienware-M11x-R1
Here are my contributions (it is nice because it's only a 1-2 minutes work that could help a lot of people and save them a lot of money (and this even if they don't an only-free distribution but any Gnu/Linux Distribution).
[GPU] nVidia Geforce 9800GT http://h-node.org/videocards/view/en/758/NVIDIA-Corporation-G92--GeForce-9800-GT---rev-a2-
[Notebook] MSI MS-1722 http://h-node.org/notebooks/view/en/757/MS-1722
[Netbook] MSI WIND U100 http://h-node.org/notebooks/view/en/756/Wind-U100
[Printer/Scanner] HP PSC 1200 http://h-node.org/printers/view/en/755/Hewlett-Packard-PSC-1200
Very true. Anything that helps people avoid non-free dependencies (and particularly those using non-free operating systems) is a good thing. People don't understand the issues with non-free software (either technical or ethical). It's particularly troubling on distributions which include non-free software. People blame “Linux” for there problems when it's the manufacturers who are the real problem. I think the free software stance is a good stance to take. I think I'd even take it a step farther though. I don't feel that we should hold free software back because some companies (NVIDIA/ATI) refuse to cooperate with the free software community. If users want to switch to free software teach them to buy free software friendly hardware. That creates a market and ultimately will solve the problems free software users (and others on GNU/Linux distributions which include non-free drivers/firmware) face.
One of the things I find humorous is Linus Torvalds stance on proprietary software. He accepts it into the kernel. I can understand his non-ethical justifications (not that I agree with them). However he then gets upset with companies like NVIDIA whom refuse to cooperate. He is acting as an enabler of these companies bad behavior.
The problem here (in my opinion), is that people don't see the ethical problem, they see a "does it work" problem. I had a conversation with a friend of mine today and the same words pop out of his mouth like anyone else that has problems (one more time, in my opinion), to see the real problem. So the discuss was around the fact that I was saying to him that I wouldn't create a proprietary software because it would remove the freedoms of the users, neither I would use a non-free software to make it just because it's a popular software for gaming production (Unity), I continue saying that I prefer something that can have bugs, but it's free (as in free speech), and gives me the freedoms I want and need, than using something closed and "non ethical" just because it's great, fun, and everyone uses.
The conversation then started to go around the fact that many people don't care about the code, and if they don't know how to edit or even understand it, and if the creator drop the project, then we will have a dead project. But once again, anyone has the freedom to take it and do whatever he wants... The best thing I can remember was when he bought Diablo III, and now he just can't sell the game because of the DRM that makes a "strong glue" between that particular game and registration code, and his own account with all the vital informations (Credit Card etc..).
This is an example of the mentality of a lot of people outside, and it works the same way for anything, software, drivers, hardware. People just feel the need for things to work, and don't care about anything else.
Around me, a lot of people started to thing that I'm stupid (literally), just because I refuse to use flash if the content they are showing me don't work with gnash, lightspark or even html5 (and I already had the beautiful occasion of laughing damn hard because an old flash video worked with gnash but not with the newest versions of Adobe Flash).
As Richard Stallman said, it is not (only) about the Software, it is about teaching people the ethical problem with proprietary software, but if the person don't want to listen, then we can't force them, it's their own life, we can just go see another person that want to learn. And works like that allows to build some great communities like this one.
It is true that we (I say we but I'm a young trisquel user with just 1 week of experience on this forum) have already found, and will find more and more people don't understanding the "ethical" problem, and starting asking on the forum for how to install X or Y proprietary software, and when the time comes we need to make them understand what "we" are doing and what "we" are fighting.
Yes RMS can be very radical, like when he says to Lunduke that he prefers that Lunduke's children start "starving" instead of him (Lunduke) starts producing proprietary software (okay very badly explained but I think you know what I'm talking about). But has he (RMS) said, it is not a reason to forget our freedom and our rights, we need to find a balance, and proprietary software, should (in my opinion), never be near that balance.
The other problem is that it's hard to understand if a Software is or is not Free, and what are the rights and the rules that he (developer or user) needs to fallow).
As for Linus Torvalds, even knowing that I really admire this man, I think that he is more from the side of the "I just want this to work nothing more nothing less", and sometimes finish by accepting everything. I think we can understand that (at some point), and this is why the Gnu Hurd kernel was created (there is even some distributions around this kernel like the ArchHurd).
So for summarize : Yes you are totally right :D
Heh, I remember when I used to be like that, where I didn't care about the operating system or software license as long as it worked. All I cared for was 'did it work, if not, how do I get it working?'. That's what I was like when I first started using Ubuntu mostly in 2009. Back then, I didn't even really know that graphical acceleration was only thanks to non-free firmware (radeon). One of the first things I would do after installing Ubuntu is install Flash, not even realizing the difference between proprietary and Free software.
In fact, the first time I learned of the whole 'Freedom vs. Proprietary' thing is when I first heard and installed Trisquel in late 2010 and wondered why there was no Flash player, besides Gnash. That's when I started to learn, and made a commitment to only use Free software from that point onwards. Yeah, I lost 3D acceleration, but oh well.
And since that day, I realize that it was the best choice I could've made. Today, I don't use Trisquel, but I do thank it for starting me out on the road to Freedom. Today, I use a system made up of fully Free software thanks to Parabola, and not only do I not have any problems with both the base system or even any applications, but it's also very, very powerful and stable. I haven't had a single problem that isn't easily fixable in months now, and I have a feeling it's because I don't have to wait for manufacturers to do a 'fix' to their software. I'm not dependant on them.
About games though, yeah, I do admit that the Free Software community has...little to offer to games when compared to proprietary titles. It's quite sad, but oh well.
And don't worry about people laughing about you for not having Adobe Flash. I've gone through the same. A few months back, I went to an online high school (that I had started when I was still using Windows...) and I asked them if they had any plans to switch from a fully Flash-based interface to something that would work with Free Software. They basically laughed at me, and months later, I dropped out (for multiple reasons, Flash being one of them) and wiped Windows off of my system on the same day.
And now, I'm trying to find ways to give back to the Free Software community.
You got it spot on. This is one of the two reasons we don't ship hardware dependent on non-free software. It's not "just" an ethical issue. There is a practical side to things too. It is unfortunate that the whole thing gets written off because people feel that it's fanaticism. It's not. There is a real world effect of non-free software on users. It's not just a development issue. "Open source" is the wrong way to go. It's not really "open" after all. It's really just free to the point it's not.
I disagree with the idea that the free software movement has nothing to offer to games or gamers. I think you'll see my point in a minute. Free software games are better in at least a few ways. They can be made to work with future versions of an operating system regardless of the original developers business model (provided it has one). There are many games that don't work today because the source code was never released.
Open source is wrong. It irritates me to no end how companies claim to be 'open source friendly' because they release a very minimal amount of source code to the community. They claim to be 'collaborating with the community', when all they're doing is making it easier to run their proprietary software using /some/ software with source available. Sometimes, the source they release might not even be under a free license, and even if it is, it's useless because it's just good for a proprietary program. That's just wrong.
I remember one time talking to somebody about 'open source'. They said that they love open source. When asked why, their reason was because they could take the code, make it better and then release it as closed source product (if the license permits) and make money off of it! Again, that's just...wrong!
I'm glad though that you have the same views that I do on 'open source'. I thought I was the only one who referred to open source as software with some source code available while the rest is proprietary, while Free Software is fully open and respects the four essential freedoms.
The big picture on open source is just terrible though, and completely ridiculous. Releasing some software's source code and keeping the rest disclosed isn't 'open' at all, and the only reason companies even release the little software that they do as 'open source' is just as a business act, so people can think they're being friendly and helping to promote software freedom, when they aren't. They're just giving you a wrapper to use for their proprietary software.
Anyway, onto free software games:
Yes, I do agree that free software games are better in the sense that they can be made to work with future versions of an operating system, and not /all/ free software games are bad. Some of them, Xonotic for example, show that the free software gaming community can really produce some amazing stuff that's comparable to proprietary, modern games. Unfortunately, those Free games that are capable of competing with proprietary, modern-day games are very little. It's sad to say it, but that's my opinion. That's probably the only thing I miss about Windows, was all the games. (I used to be a heavy gamer up until a few months ago, before the switch to full freedom). Of course, don't take that as a sign that I want to go back to Windows, because I really don't. I've gotten this far into freedom, I don't want to take a step back.
Open source is about the convenience to developers and business. It is a valid business reason to do so. Where it might actually help a competitor companies are extremely careful not to release the source code. It's why Intel releases 100% of the source and NVIDIA/AMD won't. You find that in a lot of cases its the lesser performing of the hardware which is really well supported in GNU/Linux. Humorously this frequently gives it a significant advantage from an ease of use stand point and speeds up a persons computer too. When source is shared between programs/drivers/etc the quality of the code improves and there is less demand on system resources. This is generally speaking of course.
The solution to the gaming problem (lack of quality games) is to start businesses and promote the free software business models in the game arena. It's a bit of a step to go from what I do to running a development shop that creates free software games. However I don't see any reason the same thing can't be done for games what we're doing/going to do for free software in the desktop applications area.
>Open source is wrong.
Amen! Here are a few excellent articles on the subject
* http://www.gnu.org/philosophy/open-source-misses-the-point.html
* http://mako.cc/writing/hill-when free software isnt_better.html
* http://mako.cc/copyrighteous/20120809-00
"Open source" is something most companies which love proprietary software see as something they can work with. Free software's march is inevitable so they embrace the least different form of it aka open source and try to legitimize it while completely ignoring the much more potent concepts of free software and copyleft.
My hate for proprietary software started way before I heard the words "proprietary/free/non-free/opensource software", when you start to see applications crash, with licences everywhere, and you can't replace them, this is when you start to think a little bit more about the "what's inside the box" and that's (in my opinion) the moment when you switch to Gnu/Linux and then to Only Free Distributions.
The problem with big applications and games, is that it requires a lot of people working at full time, and most of the Free projects are... well... Free as in Free Beer, and only donations don't allow the developers to work at full time on such big applications, and because of that applications like Blender, Gimp, Kdenlive start to have problems with continuing their projects as they would love to (even if the first 2 are in a better financial state that the last one).
As for games, if it's a 2D game then one person could do it (I am starting that will be Free Software (but paid) and probably I will use Kickstarter to promote that, but that's another question), but for 3D it is so much work that, or you work with applications like Blender that can easily handle everything, or you will have to work with more people in order to have something working (without the guaranty of any revenue at all).
Please Name & Shame them. They are defiantly rubbish at education.
On 22/09/12 06:38, name at domain wrote:
> And don't worry about people laughing about you for not having Adobe
> Flash. I've gone through the same. A few months back, I went to an
> online high school (that I had started when I was still using
> Windows...) and I asked them if they had any plans to switch from a
> fully Flash-based interface to something that would work with Free
> Software. They basically laughed at me, and months later, I dropped out
> (for multiple reasons, Flash being one of them) and wiped Windows off of
> my system on the same day.
See- people don't understand what it means to them if something is non-free. Even Linus doesn't seem to realize how non-free software is negatively impacting users. He goes on about other stupid design moves and fails to realize the difficulty users have with non-free software. It negatively effects less technical users the most who aren't competent at fixing, updating, and maintaining it. Non-free software is difficult. There is no question in my mind about it.
While free software can be a challenge it's not because it is free.
NVIDIA is one example of this (to get "full" support non-free drivers are required), non-HP/non-standards complaint printers are another example of this (which use non-free drivers), Realtek, Ralink, and other Wireless N USB adapters are an example of this (non-free firmware), and many many other devices.
They cause problems such that users can't utilize other features due to integration issues with free software (power management), they create upgrade headaches (system updates force you to re-install a driver), or you have to manually install a driver using some obscure and difficult procedure (open the terminal, type in commands, and so forth).
Even after all this is partially "solved" by moving to a non-free approach such that Microsoft Windows uses it still really exists. You end up with hardware that doesn't work any more because the drivers are no longer supported by the manufacturer. An upgrade breaks support. Except in GNU/Linux and purely free software environments those drivers are not dependent on the manufacturer.
Just to correct an historical mistake: GNU Hurd was in no way developed in reaction to Linux. Its development actually started in 1990, i.e., before Linus Torvalds started Linux. Actually, the effort on the Hurd was rapidly diminished because, after being released under the GNU GPL, Linux was filling the missing piece (the kernel) of the GNU operating system. There was no pressing need for another kernel (whose development is hard because a micro-kernel architecture was chosen).
Its so good to notice that my little post raised so much discussion. Getting back to the topic: h-node. I am currently finishing my studies by doing my practical training period in a small computer shop repairing computers. This gives me an excelent chance to contribute to h-node.
When I run into a machine which is in working condition, I test for h-node. I can usually find a couple of machines per day to test. My practical training period lasts until the New Year, so I expect to be able to make a huge contribution in that time.
@ Chris:
---
You mentioned that people tend to use h-node only to check if a particular piece of hardware works and then make their purchasing decision depending on that and thus h-node does not really tell the users clearly which companies are really pro free software and which are not.
Quote:
"For example it makes a lot more sense to go with an Atheros wireless PCIe chipset over the equivalent Broadcom chipset.
While both are freedom friendly Broadcom has not put forth the same level of support for free software. Until recently Braodcom didn't even offer a driver, documentation, or any cooperation at all on GNU/Linux."
End quote
You can report this as an issue in h-node's issue page. You can also write a wiki article about this subject to h-node. Your contribution would be most welcome since you have a lot of insight to this matter as an entrepreneur in the field.
---
Awesome! Thank you for your contribution!
I definitely should. If I had the time to write a book I probably would.
I'm so busy trying to scale a small operation into something bigger though it isn't like to happen.
I think a better solution is to advise people that the information can easily steer people in less desirable direction and that it would be better to support companies like ours (libre.thinkpenguin.com) and any other companies who would take the same approach to free software.
I don't think a monopoly is a wise thing. I do think it's better to throw peoples money into a funnel though. Combined the resources will do a lot more good. This can be seen with the USB N wireless adapters we are working on. And that is just because there are zero good USB adapters readily on the market any more that a decent amount of it is being funneled our way and resulting in change for the better.
Anyway- despite my criticism h-node is helpful compared to having nothing to go on. At least with it you don't have to go through a dozen different cards/adapters/printers/etc before eventually coming to something that at least partially works. Which is an improvement over having no resources at all. However this isn't a workable solution for anything other than the minority who are already on GNU/Linux.
I was wondering if I would suggest a system on h-node to recognize truly free software frendly vendors from the rest. Let's assume we have a list of those vendors. Then it would be relatively easy to add some kind of mark or message to every device's page from that vendor. It would also be easy to use this recommendation as a filter when searching devices from the h-node's database. This way people could not only find devices that work more easily but their support (purchases) would get directed towards free software frendly vendors more easily too.
The problem is that h-node does not have such a list. Who does? FSF? It would be good if some third party like FSF would maintain such a list which h-node could then rely on (like it does now with FSF's list of fully free GNU/Linux distros). It shouldn't be h-node's responcibility to maintain that list. Do you know if FSF or anyone else have such a list?
@ SirGrant: Thanks!
This is the list of wifi card chipset vendors that provide only ones
working with free software:
(empty list)
I think it would be similar for other devices. Seems easy to support.
A random card with an Atheros, Ralink or Realtek chipset might require
nonfree firmware, it's similar with devices having Intel GPUs (some are
PowerVR). Some Broadcom wifi and Nvidia graphics cards work without
nonfree software while the vendors don't support it. It's simpler with
Intel wifi or AMD GPUs, none work without nonfree software.
Or do we use "vendor" in a different meaning than on h-node?
ThinkPenguin uses only chipsets that have free software drivers available. On the retail side I think it's more then possible to point to companies selling to end users. The problem is there aren't any companies besides ThinkPenguin selling only freedom friendly hardware.
As far as the chipset companies go Atheros does a good job with some although not all chipsets. They do a good job with PCI/PCIe chipsets. They have / are cooperating with USB N chipsets. I'm hoping to get them to do the same for Bluetooth PCIe.
Intel does a good job with graphics although is hostile to a free BIOS.
AMD is hostile period. They do like the public relations benefits of supporting "open source" though.
NVIDIA has been hostile. That might be changing or it might not be. AMD pulled a PR move by releasing some code/specifications and now everybody thinks they are great despite failing miserably to offer something that is freedom friendly.
HP is pretty much the only printer manufacturer which offers good documentation and support to the development of a free software driver for a respectable portion of their printers. Canon, Lexmark, Samsung and others which have provided drivers in some cases for some markets have failed to deliver. They were non-free and a real pain for the users who didn't know or didn't care about freedom.
The webcam market is pretty straight forward. My understanding is there isn't anything complicated about writing such drivers and pretty much everything has a driver. That isn't to say every device works though. The credit to this goes completely to the free software community which developed said drivers I believe. I don't know if any webcam chipset companies provided support.
dial-up modems are interesting. There is a standard driver that supports some USB modems. Any card that works with this driver is probably the one you want. I don't know who made that driver possible, how it came about, or if one of these chipset companies helped.
I'm not sure about USB ethernet chipsets. I believe that all the 10/100's are all free and the code probably written by the chipset companies. The 10/100/1000 adapters are not all freedom friendly. Nor are all the PCI/PCIe/laptop 10/100/1000 chipsets.
That is almost all of the info I have on this subject. At least of the hardware we sell. There are some others we have investigated a little like TV cards and HD decoders.
Does the ThinkPenguin computers also use a free bios (coreboot for instance) ?
No. It's not currently possible. Even with a free BIOS it would not be 100% free due to other microcode. Porting a BIOS is also a non-trivial task. It is too costly to be feasible at this time. It's not likely to happen any time soon either for various reasons. Intel has been uncooperative in supporting coreboot developers. Amd does support coreboot although doesn't support it's graphics chipsets.
Systems with AMD CPUs on the laptop front ship with AMD graphics or NVIDIA graphics.
Systems with Intel CPUs can ship with Intel, AMD, or NVIDIA graphics.
Intels CPUs are integrated in the CPU and therefore Intel graphics can't be used with AMD CPUs.
Well, h-node wasn't really made for this purpose. I'm not saying it's a bad idea or that it wouldn't be welcomed. Part of the problem I feel is that companies also purport to be freedom respecting and aren't. At what point do you classify a company as being freedom respecting? Is it when they only ship chipsets which have free software drivers? Or do you do it when only use chipsets which are supported in some way by the chipset vendor? For example HP offers that although NVIDIA does not.
NVIDIA's stance on free software is hostile. They've taken an approach that hinders development of a free driver. However there is a free software driver for older chipsets that was reverse engineered. Do we say that is OK to sell such hardware or not? libre.thinkpenguin.com offers one card with an NVIDIA graphics chipset that is well supported by the free nouveau driver. I'm not a fan of NVIDIA over this. However we are offering one card. We won't do that for laptops and desktops for various reasons. For one the newer graphics chipsets are not sufficiently (and at all) supported by any free driver.
FSF has a program which we will be participating in. It's not so much a list as an endorsement.
I'm kind of of the perspective that we need to support whomever is doing the best job for any particular category of hardwre. If that is Atheros then we should support Atheros. If that is HP we should support HP. If that is Intel we should support Intel.
Now these companies don't provide freedom friendly chipsets everywhere. They only tend to do it where someone has convinced hire ups that it would be beneficial to the company for them to do so.
There are problems with this though. For instance Intel graphics chips are tied to Intel CPUs. Intel has been uncooperative with the coreboot (a free BIOS project) though. Do we classify them as hostile or not? AMD has also been hostile though. AMD does work with the coreboot project although has not resolved the issue of shipping a partially free driver that is dependent on a non-free component.
Then if you get to the retail side of things there really isn't anybody out there selling freedom friendly hardware except us. Sure- there are companies claiming to do so. But what they actually ship is hardware dependent on non-free drivers/firmware. Some of it you won't even be able to make work with a free distribution because of digital restrictions in these systems.
The reaction of the people running them is not good. They give poor excuses as to why they can't ship freedom friendly hardware rather than fixing the problem (which is not technically an issue for them).
Quote:
---
I'm kind of of the perspective that we need to support whomever is doing the best job for any particular category of hardwre.
---
End quote
I agree. If there isn't any hardware vendor out there which would provide only and completely free software friendly hardware then we have to support those who come closest to that and hope that, in time, our support guides them to increase their friendliness towards us.
I think we can recommend them on h-node like this:
---
Recommended hardware vendors:
This page lists those hardware vendors who are free software friendly.
The following hardware vendors provide only completely free software friendly hardware:
* none
The following hardware vendors are quite free software friendly, but not always:
* HP
* Intel
* Atheros
---
Something like this I think would be good. I think it has to be very carefully worded though as you can't blatantly recommend HP, Atheros, or any other company. We're the only ones which are exclusively shipping freedom friendly hardware.
As the example I gave earlier HP is the company to go to when buying a printer. However it's not possible to just go into a store and pick up any old HP printer. Your going to have to do a lot of digging to determine which model and which version of a model even is compatible. It's documented so those who know can do it. However that doesn't mean it is going to be easy to get that printer.
I made the decision earlier today to get 3 years worth of stock for a particular HP printer for instance. This printer is no longer available and the newer models are not supported out of the box on the LTS releases. Users of Trisquel 6 for instance won't be able to find a printer that works out of the box anywhere as HP's entire lineup which has support in LTS is discontinued. To get around that we're stocking up.
You Chris are probably one of the most knowledgeable people when it comes to hardware free software compatibility. I think you might want to talk to the h-node guys to improve the site further. I've found them nice to deal with and open to new ideas. It would benefit us all.
Hi, this is a very interesting topic.. H-node is born with the aim of collecting hardware and many steps have been carried out in order to make its information useful for the free software community. Many other steps have to be done.
We should distinguish between sellers and vendors. At h-node.org we are listing vendors, not sellers. So for example AMD is a vendor, ThinkPenguin.com is a seller
So there are two things that could be done:
1 - find a way to rate the vendors, as suggested by lembas inside the h-node.org issue page
2 - indicate where the devices can be bought (link to the sellers)
Point 1 is very critical: we could both use a statistical approach by counting the number of working device (that we have at h-node with the same vendor) as a percentage of the total number of inserted devices and/or use some arbitrary parameters chosen by the project (do that vendor provide assistance to free software users? , do that vendor provide free drivers?, do that vendor avoid to use restricted boot?, and so on).
We should be able to produce some kind of h-node index for each vendor. But as you can see it could be very arbitrary. I think that the easiest way would be to manually rate the vendor. For example, Apple should be marked as hostile even if, for example, some of their devices could work with free software.
Point 2 is easier to solve: we could create a page that lists all the sellers that we think are good choices to buy a computer that works with free software (ThinkPenguin.com could be a good choice). But the sellers should be certified in some way. I think that the sellers should be certified by FSF (also because h-node is now a FSF activity). I don't know if it could be feasible.
In principle I agree with the fact that who sells computers is the person who knows best these problems, but even if some sellers (such as thinkpenguin) seems to be very good choices, most of the sellers aren't reliable. So I think we should let the users to specify all the information as best as they can, even if in some cases (for some devices) the result is still a bit confusing (i.e., it is still not so simple to choose a device on h-node and then go to a shop and buy that device).
But we can start thinking a way to choose, among the collected information at h-node, a way to select and suggest those devices that we know are working and we know that can be easily found (without the risk to confuse it with a similar device with a different chipset). This is a feature that have been discussed some time ago, a way to suggest a device, where the user can give a rank (positive or negative)
At the French level, and considering more legislative concerns (whether the price of the included software is provided, whether their licenses can be read before the purchase, etc.), these two sites, maintained by a GNU/Linux user association, exist (if you do not read French, you can try an automatic translation):
- For point 1: http://bons-constructeurs-ordinateurs.info/
- For point 2: http://bons-vendeurs-ordinateurs.info/
By the way, ThinkPenguin is not listed in the second site. It should be on the top. I guess Chris will want to write to name at domain (I can help with the French but English should be fine I guess).
I'm not going to write to them at the moment.
:) Our support beyond English is very limited to non-existent at the moment. Feel free to do so though if you like. It's a work in progress. The plan of action is to get the new site up (had an Oct 1st goal for the finishing the code portion- it's being extended though due to complications). French though is one of the languages we should be able to support fairly easily across our entire product line up once the people and systems are in place to do so. French, English, & a few others. I'm hoping to eventually be able to support nordic countries better.
While it might be difficult to impossible to find some of our products elsewhere some of them are manufactured for us specifically. We are working with a distributor for instance. Although you won't necessarily find any of the products on store shelves at the moment. The customers are larger corporations.
I think we have to forget statistical methods while our sample is still relatively small and opt for a general description (and ratings for easy access to h-node users) of the vendors. Many vendors will get a bad score but that unfortunately only reflects reality. This is something we should not be too scared of.
I'm a big believer in open wiki constructs. This is the kind of web 2.0 I can believe in!
One more thing I think could help is getting the h-client included into Trisquel 6.0 by default. That would help increase the participation in and awareness of h-node. Ditto every other free distro!
Some time ago when I was younger and (even) less well informed I helped a friend install Trisquel on his HP laptop. I too have a HP, inherited, not selected. Turned out his wifi wouldn't work and for some strange reason we couldn't get the native resolution of his screen... He remarked that I had gotten lucky with my hardware because my wifi and my screen worked. Indeed. But we shouldn't have to rely on luck. That's why there's h-node!
> One more thing I think could help is getting the h-client included
> into Trisquel 6.0 by default. That would help increase the
> participation in and awareness of h-node. Ditto every other free
> distro!
See http://h-node.org/wiki/page/en/ToDo#h-client for some features that
might be useful. I think we need translations before adding it to
Trisquel and maybe some code changes.
:)
h-node is luck though. ThinkPenguin is not luck. We're shipping back 35 printers because of an issue tomorrow. This is for a printer we have been selling for at least a year. Trust me. It's luck when you get hardware that is what you expected based on a model number.
If other sellers participated in a program to ship freedom friendly hardware it also wouldn't be luck any more when you bought from them and got what you expected.
FSF has a program to endorse hardware. We're going to be involved in that and it would be good if others did too. I doubt there will be many putting up with the trouble for what they see and a insignificant customer base though. That's reality.
I like the idea of h-node. I just don't think relying on "manufacturers" which swap out chipsets on a whim and don't update the model numbers is doing a good thing. It is doing a disservice to free software and GNU/Linux users in general. And particularly those who are actively engaged in supporting the companies whom are doing this. If you are a seller of "Linux" hardware or to a GNU/Linux user base you should feel obligated to solve these issues.
The sellers should know better than to sell hardware with non-free dependencies, the user base should be informed, and users at h-node should be warned of the reliability of the information available. I should also point out this isn't just true of h-node. It's true of every similar type of database on this subject. H-node is nothing new. It's just slightly different and slightly better in that it doesn't include hardware with non-free software dependencies (I should say tries to avoid since we never know waht is actually going to be in that Acer Laptop model #3KB2A or USB Adapter AR712S).
http://en.opensuse.org/Portal:Hardware
https://hardware.redhat.com/
http://community.linuxmint.com/hardware/search
We shouldn't need a different database for each and every distribution under the sun either. But ohh well.
I don't feel my time is worth focusing on this issue. I think we can solve this issue best by focusing on improving the availability of the hardware rather than making it slightly easier to find potentially compatible hardware that is still hostile to users.
> for some strange reason we couldn't get the native resolution of his screen
Sounds familiar. It took me hours to get xorg.conf working on my gNewSense 2.3 setup, on a Dell laptop.
How long ago was that? Because those issues don't occur for me on newer operating systems like Trisquel 5.0+. Thankfully a lot of work has been done and some of those issues are disappearing.
But definitely agreed on H-Node. It's a great database, and I'll use it when looking for my next laptop (which won't be for a few years).
So perhaps what we should start doing with hardware is getting it labeled properly, chipsets and all. Sounds like your average "consumer protection" thing. And since this is not GMO food we're talking here about it might actually happen.
>How long ago was that? Because those issues don't occur for me on newer operating systems like Trisquel 5.0+. Thankfully a lot of work has been done and some of those issues are disappearing.
This silliness happened on 5.0 which was then upgraded to 5.5 with the issue remaining. However I had "only" a few hours time poking at the thing... That was on some about 5 years old nvidia card.
We should rate vendors both in general and in detail. For example, we can give a vendor a general grade (x/10) and then more specified grades on different subjects like attitude, code contributions, reliability, quality of products etc.
Each user could give these grades and average grades would be calculated:
HP ..................... x/10 (votes: xxx)
attitude ......... x/10
reliability ...... x/10
c.contributions .. x/10
etc.
We should also list good sellers like ThinkPenguin.com and give them similar grades:
Seller X ............... x/10 (votes: xxx)
quality .......... x/10
freedom
frienliness ...... x/10
prizes ........... x/10
etc.
hmm prizes. We'll have to come up with something.
:)
I just checked my hardware on h-node, even though I know it all works with free software & deblobbed kernel. The output was as follows:
--------------------------
The following devices has not been found in the database:
can you please insert them?
videocard - Cedar PRO [Radeon HD 5450/6350]
device type: VGA compatible controller
vendor: Advanced Micro Devices [AMD] nee ATI
VendorID:ProductID code of the device: 1002:68f9
--------------------------
After clicking 'insert' it takes me to a page that lists ALOT of hardware (is like 4 pages long), and I have no idea what to do at that point. All I know is it works just fine without the non-free blob. I have no idea why h-node doesn't 'insert' the information automatically. Once I click 'insert,' it should just do it without any further input from me.
I built this computer last year and did my own research before deciding on the hardware, because I knew I wanted 100% free system. I didn't check h-node though. My old motherboard had a nVidia card and I used the closed drivers for 5 years. Honestly, it got old having to rebuild the driver every time I updated the kernel.
What ticked me off most about it is, once I made the decision to switch to the nouveau drivers, most distros would leave only a black screen at install. That ticked me off, because although I had decided to go will an all free system, I couldn't, because the distribution packagers were not configuring things correctly at install. I believe now the issue had to due with KMS.
So when I had this box built, I decided to leave nVidia completely and bought the radeon, which works perfectly, because AMD has been working for some time to improve the free drivers. (I know, I know. AMD gives specs here and there, but not enough info to really do it right. At least they are doing something though, unlike nVidia.)
'm no fan of NVIDIA as they have not been good to GNU/Linux users. They cause all sorts of trouble. The reason older cards have support with the free driver is because of reverse engineering. The 9500GT is the best supported card using the free driver. It's good enough to get into the mainline kernel. That is impressive.,
,
The only graphics I'd recommend is Intel. They do a good job of supporting free software (graphics chipsets ONLY and only ones developed in house).,
,
AMD does a horrible job in the graphics department. You might say it was better than NVIDIA which offered no support, source code, or specifications although AMD doesn't provide enough information for any support on free systems either. It's all but worthless as far as I'm concerned. Camouflage for there proprietary garbage.
Hi everyone,
I've been learning a lot about Trisquel lately, mostly because of h-node and what a great resource it is. I've actually been working the past few days on a crazy idea to solve the problem that you already brought up:
"We shouldn't need a different database for each and every distribution under the sun either. But ohh well."
I've been building a unified database of test and certification information, and I'm using Trisquel and Ubuntu as guinea pigs for it, because you've made the data so easy to use. My goal is to generate money for the distributions through the site. It's very rough right now, but you can see there's a lot of data in it. I'm using only the "A-platinum" rated devices, and for now starting with laptops.
Here's an example page combining data from both distributions.
http://www.linuxshopper.com/devices/asus-eee-pc-900
If I understand what you are trying to do this is not a solution. Merging the info does not solve the problem. The only thing that would solve the problem is to eliminate devices dependent on non-free bits. Then certify those devices and sell them under a particular brand/certificate so users can pick out devices that won't become an issue and distributions can properly support.
Even though we have people using Ubuntu and Trisquel who have said the Eee PC 900 works there is no guarantee (and even if it was true of this device it isn't of others) that all Eee PC 900 systems are the same. Just because a dozen people from a few distributions have working Eee PC 900s does not guarantee it'll necessarily be true of all Eee PC 900 systems. Model numbers frequently cause issues as people assume them to define a particular version when in reality they do not. Making purchasing decisions based on them does not work well. You don't have enough data points to rule out a device as having multiple versions or not either. It's never going to happen. Even if you can say that the Eee PC 900 today has only one version how do you guarantee that the people who buy it in the future won't get something else? You can't do that without certification and assistance from the manufacturers.
Yes, a site like I'm making is only as good as the data coming in to it. But this testing data is what people are using right now as part of their research when buying computers. I'm focussing on making it much more accessible, while linking back to the original sources for users to verify for themselves the meaning of "certified". It would have saved me a lot of time, had a site like this been in place when I was looking for my last computer.
By the way - I didn't see Think Penguin systems in the database, yet they are compatible. Do you plan on entering data there for them?
- Login o registrati per inviare commenti