Vendors Paying Lip Service To Linux Support?

NetJunkie asks: "Most people are familiar with the Microsoft and Novell 'certified' logos on software and hardware. If something is shown to actually not be compatible with a standard install they can lose that logo. But what about Linux? I bought an IO Gear KVM/USB switch with the Tux logo on it. When it didn't work correctly switching consoles (the mouse would stop working under Linux until I reloaded mousedev) I called tech support and was told '...we don't really support Linux and it hasn't been tested very well'. They couldn't even give me a tested configuration that was proven to work, and I already run a pretty stock configuration. So what do we do?" We fix the problem. If their support doesn't help, maybe some kind coder out there has been able to solve the problems and can lend a helping hand. What other things can we do to ensure that vendors who say they "support Linux" really make good on their claims?

Comments

This is unavoidable

[cowscows]

This is part of the commercialization of linux, and it's unavoidable. You're never going to run out of some people trying to make a quick buck at the expense of others. I guess another type of information that you'll need to share is which vendors are trying to screw you over and how.

And as linux becomes easier to use, and more accessible to the average user, this will get worse. You've got the advantage of most of the current linux user base being able to tell that they're being screwed over. I guess just spread the word of who sucks as fast as you can.

Non-partial testing company needed

[linuxonceleron]

We need some kind of non-partial company to test for linux support. Of course this will cost money for the producers of the hardware, but there's no way you could do it for free (short of giving the hardware away to linux users for testing :) ) A "Tested by Tux Labs" logo on the package would be what we would look for on packaging. It would also make things easier than having 5 redhat/suse/mandrake/debian/slackware ready logos on the box. This is an open space in the Linux business market that I'm sure someone will fill shortly as Linux becomes more mainstream.

Re:Non-partial testing company needed

[fatphil]

One problem is the following syllogism:
a) a testing lab won't exist if it loses money
b) a testing lab wil try to make money
c) there will be 'competition' in the 'testing-labs' market
d) You'll not know one 'works in linux' logo from the next.

I think it's probably better for it to remain in the hands of the distributions. Their names and their support structures are already in place.

Until then the best thing we can do is to
a) write to mailing lists about incompatabilities found, keep everyone informed; then
b) vote with your wallet. Don't buy those that lie or only pay lip-service to their support claims.

FatPhil

duh er...

[B1ood]

don't buy the product. if you bought it, take it back. tell your friends not to buy it, they tell their friends, etc.

sure, someone on a mission could start giving personal endorsements, but it all boils down to "who you trust?" do your homework on a product, and if you get burned, prevent someone else from doing the same.

B1ood

Re:linux stickers

[linuxonceleron]

linuxmall and copyleft have linux stickers that are made for your computer case, I'd recommend getting the big sticker sheet from linuxmall, its like $2 and comes with more stickers than most people have computers.

No single controlling entity = no certification

[dustpuppy]

Novell and MS have their certification logos and they can be trusted. Why? Because there is a controlling entity (ie MS or Novell). If a vendor misuses the logo, then bang, there goes the certification logo.

But with Linux, there is no real controlling entity - Linux's strength is in it's diversity and dispersed control. But this also makes it difficult to have any sort of certification that can be enforced. If there is no one entity saying what constitutes certification (and no entity to revoke certification), then all we can rely on is the reputation of the vendor.

The only way to fix it (as far as I can see) is that we need a single entity to test and validate Linux certification. It think that something like this will start eventually.

The real danger comes when, becuase of Linux's open source nature, we have multiple certification standards which will lead to confusion and will ultimately prove detrimental to Linux as a whole.

Re:No single controlling entity = no certification

[PsychoSpunk]

But with Linux, there is no real controlling entity - Linux's strength is in it's diversity and dispersed control. But this also makes it difficult to have any sort of certification that can be enforced. If there is no one entity saying what constitutes certification (and no entity to revoke certification), then all we can rely on is the reputation of the vendor.

Someone has to own the rights to Tux. And someone very prominent is the owner of a little known word: 'linux'. I believe his name is Linus.

What I'm getting at is that yes, in the sense of Linux (the OS), there is a controlling entity, and it could be said that he and the lieutenants could authorize a group to do said logo control. That way end users (read newbies) don't get burned and go back to their previous OS. (Note, not all end users are newbies, but I doubt a long time linux user would give up that easily.)

great topic.

[Bad_CRC]

This is one of the things about Linux that people don't tend to talk about too much.

Almost everything in linux is sold or marketed as "unsupported" from the corportations and hardware manufacturers who market them.

Now, we can't try to go around demanding full support at this point, as linux is still small enough that these companies would probably be more profitable by skipping linux altogether than by investing in full support for linux drivers and support staff. Especially with the incredibly high number of distributions and variations out there.

A type of certification would be great, but implementation and enforcement would be a nightmare.

We can't demand more support, but if we settle for no support, that's all we will ever get. No easy answers, other than the only one which makes sense at this point... "wait till we get a bit more marketshare"

________
1995: Microsoft - "Resistance is futile"

Market is too broken - no central linux regulator

[jonnythan]

There's an obvious problem with this: there is no central authority who has the power to hand these guys a cease-and-desist order or risk prosecution.

With MS and others, they have the power to actually tell those companies "take off our certified logo or we'll sue you." If someone indicates linux compatibility and the product doesn't quite work and you can't get support, there's no company or organization you can go to and say "look what they did! make them take it off!"

I think it would be very beneficial for linux in general if some of the large, more trusted names in Linux got together to organize a Linux Certification system. The whole linux name would gain some credibility if Red Hat, Corel, Mandrake, Suse, VA Linux, etc., formed something by which they would have the power to give and take "Linux Certified" stickers or something and give out "Linux Certified System Administrator" certifications.

This way, people might eventually recognize a little penguin means nothing, but the nice pretty logo branded with Red Hat and Debian logos actually means something. Those friends who are studying for their LCSA's are doing something useful and profitable too.

Until we get enough linux companies together to agree on such a certification system, it will remain totally out of our reach to make sure that companies who claim to "Support Linux" really do.

Tux logo

[Todd Knarr]

In this case, you'd have to go to whoever owns the rights to Tux. They could go to that company and tell them to stop using Tux to imply Linux compatibility or be sued for trademark infringement, and back it up. That's about it. Distro brands like RedHat or Debian could enforce this on use of their logos as well, eg. Debian could keep companies from using the Debian logo on their products unless they satisfied Debian that their products worked to Debian's satisfaction with Linux.

what i'm thinking...

[fluxrad]

is that we should have linus, or more realistically, a third party take charge of the regulations. If your software doesn't run reasonably well on the majority of the default linux installs, then you don't get the seal of approval.

There are, however, two major problems with linux compatibility certifications. 1: Linux is, by it's nature, a completely free and open environment. This means that no two people really have the "same" linux install. Everyone's got their install tailored to their individual needs. this makes it really hard to say "yeah - this product will work for you." There's no way in hell a company can test it's software on thousands of different configurations. It's just not cost effective

2: Linux doesn't really hold the same hand that Windows does. To a company, getting your Windows cert yanked is a MAJOR drawback - they'll spend the bucks to keep that logo - and to support it. Most companies (note: most, not all) don't give a flying fuck what happens with linux because they don't have a major investment. Linux, for companies like Corel, and certainly Adobe is just that "other" release that they're trying to pick up on. They don't really care if some organization or person says that it's not quite good enough. When linux gains signifigantly more market share, that'll change.

I suppose the best thing to do right now would to just have a sort of review repository that rates the software as far as how well it works with any and all types of linux. This wouldn't really be a seal of approval, but it would certainly give software manufacturers and consumers a benchmark at least. Other than that, the best thing to do at this point is to get more people using linux, then the software manufac's will realize that people who use linux actually DO vote with their checkbooks.


FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network

Standards.

[Signal 11]

We already have standards. You can get the latest version by typing "vi /usr/doc/HOWTO/Hardware-HOWTO". This isn't a joke, I'm dead serious. The linux standard is the kernel, so the sooner we convince manufacturers to provide modules for the KERNEL for their devices, the better we'll be. It could be worse.. they could be creating must-run-as-root daemons with closed-source APIs. I say we formalize an existing standard.. afterall, it's worked quite well so far.

Re:Standards.

[acidrain]

Too bad the Hardware-HOWTO is so far out of date. Even the kernel docs lag on hardware support. To quote Linus on the 2.4-test1 release:

...we'll find bugs. And bad behaviour. And wonderful features which we'll document some day.

Re:Standards.

[kkeller]

You can get the latest version by typing "vi /usr/doc/HOWTO/Hardware-HOWTO".

This doesn't help with closed-source software (like HP's WebAdmin software for its printers) which is advertised as "Linux-compatible" but is really "RedHat-compatible". And no, saying "don't use closed-source software" isn't always an option here, either.

and no standards

[ArchieBunker]

Novell is novell and windows is windows, but are you running redhat or slackware? X or plain console? Kernel 1.x or 2.0 or 2.2? Linux changes daily so their "test configuration" would be outdated by the time the product made it to stores.

KVM Switches & linux.

[Hardware_Bob]

The new Cybex Switches all have a linux logo on it, which says linux and then in small print "linux-tested.com" This site provides linux testing and psuedo-certification.

This shows that:

• Hardware Certification is important to linux
• Cybex (and other companies) who get their hardware "certified" or at least tested, recognise linux.
• three things in a list look much more credible than two.

I use a cybex SV-4 on 3 linux boxes daily, using an intellimouse, and have never had a problem.
--------

Re:KVM Switches & linux.

[alhaz]

I wrote the original testing specs used by KeyLabs (the linux-tested.com people) last spring. I can attest to the fact that at least when i was in charge, this was indeed a meaningful set of tests. I handed the project off to another tech (a rabid linux user as well) in june of 99 and left it in good hands.

KeyLabs definately isn't the only shop offering testing. Linuxcare used to, maybe they still do. I don't work for KeyLabs anymore, and I don't know if James is still the project lead for the linux project. What I do know about KeyLabs is that the culture among the techs is that it is more fulfilling to *fail* a device than to go through days of boring testing and just say it passed.

I took *Pride* in being a major speedbump in marketing initiatives when something wasn't up to par. When you have to spend five hours pressing every key on a keyboard over and over with different modifiers, in a SPECIFIC SEQUENCE, it's the only joy possible!

All that being said - the issue here is probably that the mouse you're using recieves some initialization from the driver (to set up a wheel, or similar) and that the KVM is inadvertently causing the mouse to be reset when you switch between systems.

Belkin used to haven an issue like this with XWindows, there was aparantly a workaround until they fixed the hardware, for the life of me I don't remember what it was, but it's probably on belkin's support site.

Re:KVM Switches & linux.

[alhaz]

Forgot to mention - *Could* also be that your system is a Micron or something, and the PS/2 ports are not PC98 compliant - that is, they don't support hot-re-plugging of mice.

Try just hot-re-plugging the mouse (pull the cable, wait a minute, plug it back in) without the KVM in series. If that doesn't work either, the problem is both that your motherboard isn't very smart and that your kvm is pretty braindead.

Re:There's no certification

[acidrain]

I'd go with the first suggestion. A few other people suggested distro's here as well, but that really just means Red Hat doesn't it? If not then it means anyone with a cd burner or a static IP. I think Linus with his main kernel source is the way to go. He already trademarked Linux, why not make a "Works with Linux" logo and restrict it's use to those who are supported by the main kernel tree. This would be another incentive for hw companies to write drivers. Besides all the PC vendors would jump on it, I'd love to see it next to the intel sticker in the store, and we could all print one to put on our boxes. Hmm, and have logo contest to allow the communinty to participate... yes... I think I should be mailing Linus instead.

Re:There's no certification

[the_B0fh]

Umm... maybe it's just me, but when it says that it works with linux, and it doesn't, welp, it just seems like the Better Business Bureau might want to talk a little with them. And if you had it mailed to you, welp, the Post Office has this thing about mail fraud (find out all about it at www.usps.gov). See, there are all kinds of consumer protection laws out there, even the state attorney general may get interested if enough people complain. And there are also laws against false advertising.

the_B0fh

Vendor collaboration on certification is the way!!

[dustpuppy]

It seems to me that it is in the interests of the major Linux vendors to form a joint business which will provide Linux certification.

That way they share the cost between them (minimising individual cost), there is inherent value in the certification because of the vendor reputations and market share, and it would promote greater consumer confidence in using Linux, and in the vendor specific distributions of Linux in particular.

As far as I can see it is a win-win situation.

Before compaining, make sure you're not the problm

[kimo_sabe]

That your KVM/USB problems sounds very much like an old bug in mousedev.c itself. I've seen that bug in action myself. The problem was that mousedev.c wouldn't notice new devices when they were plugged in. At least while the mixer device was opened.
There is no problem with them saying their switch works with Linux, or anything else for that matter, as all it should be doing is changing the electrical connections around. Beyond that it's up to the host systems to properly handle the situaition. With USB as soon as that electrical connection is gone, it's gone. Just like pulling the cord out of the plug. Switching back is just the reverse, plugging your device back into your computer.
Also, I havn't seen any recent trafic concering USB KVM's on linux-usb. Nor in my searches of linux-usb-user and linux-usb-devel on sourceforge.

- kimo_sabe

Collaborative vendor certification is the way!

[dustpuppy]

It seems to me that it is in the interests of the major Linux vendors to form a joint business which will provide Linux certification.

That way they share the cost between them (minimising individual cost), there is inherent value in the certification because of the vendor reputations and market share, and it would promote greater consumer confidence in using Linux, and in the vendor specific distributions of Linux in particular.

As far as I can see it is a win-win situation.

Re:Collaborative vendor certification is the way!

[Azog]

I second the motion!

Besides certifying hardware, there's all sorts of other things that such a joint business could do... sponsor benchmarks, pay for proper documentation of the kernel internals, negotiate with hardware vendors, etc.

Also, it would provide a unified "Linux" company that could help provide some focus for the rest of the business world when they deal with the Linux community. Right now, Red Hat seems to get that honor. And, no disrespect intended to Red Hat, but they aren't Linux, even though people talk about "Linux 6.2"...

Companies that ought to sponsor such a venture, each putting up a small amount of money:

• Red Hat
• Mandrake
• VA Linux
• Andover
• SUSE
• [your favorite distribution]
• IBM
  
• others?

Non profit orgs like Debian should get representation for free. RMS, ESR, and other deserving people should be involved in some way that isn't a burden for them, and helps keep the thing rolling in the right direction.

The company could make money by trademarking a logo and licensing it to hardware manufacturers who pass the Linux compatibility test.

This would be a good thing!
Torrey Hoffman (Azog)

Why BSD uptimes are longer

[yerricde]

BSD has much longer average uptimes than Linux

Only because BSD systems are often dedicated workstations or servers. GNU/Linux has a higher fraction of users who dual-boot because they can't afford vmware to run their precious Windows games.

A Lesson In How To Use Market Forces

[Hrunting]

I recently was in the market for a new cheapie ethernet card. I'm not talking about anything fancy here, just a little 10Mb card that I could stick in my Linux box for use with a new ISDN adapter. I went to Best Buy and bought a NetGear FA311. The side of the box said 'Supported operating systems' and listed underneath that 'Linux'.

Inside, I was surprised to find on the driver diskette actual drivers, both in compiled module and source code forms. So I followed their instructions for installing the module with my 2.2.16 kernel. The module wouldn't load. No doubt it had been compiled for something else. So I followed the instructions for compiling in their module directly into my kernel. They didn't do that. They were completely wrong. So I messed around with the kernel and driver source to get the card into the configuration routine and successfully compiled it in. On bootup, the system crashed. So I contacted NetGear technical support. I received the Old Faithful of tech responses, "We don't support Linux, but we have a text file that will help you." I followed their text file to the letter (basically, compile in support for the DEC Tulip driver) and still no go. I then replied back to the NetGear guy, "Thanks but no thanks, I'm taking my business elsewhere." I then took the card back to Best Buy where I received a full refund for a 'broken' card. With my refund, I went to CompUSA (so shoot me) and bought an even cheaper SMC card (less than half the price, and the first one cost $24) that worked like a charm.

So what do you do when a company says they support Linux and then you find out they don't? Take your business elsewhere. Even in my story I spent too much time fiddling with that NetGear, all because I didn't feel like driving back out to Best Buy. From reading newsgroup posts after the fact, I discovered that NetGear's even worse than I suspected. Apparently, their 'drivers' are actually the same thing as the standard DEC Tulip drivers rebranded, without proper crediting, and they did something that prevents the driver/card combo from working on a majority of systems, something they refuse to acknowledge.

So what do you do if a company 'supports' Linux and really doesn't? Take your piece of hardware back and say it's defective. For all intents and purposes it is. For part of your payment, you're getting support and if you're not getting support, you have a defective product. Let market forces sort out the rest.

Re:A Lesson In How To Use Market Forces

[Zigg]

You know, I had a similar experience back when I used to use Linux. I want to float an idea here. This particular situation might be more Linux's fault than the vendor's.

Linux has had a rather sad history of making incompatible changes from version to version just for the sake of making changes. It wouldn't even be so bad if you could just recompile and move on -- but in a disturbing number of situations, you can't even do that.

I will add one thing, too, at the risk of being labeled a troll: I hear cheers from the Linux-using populace when hardware vendors grumble about having to rewrite their drivers for Win2k. I hear angry self-righteous posturing when hardware vendors don't even know (and whose fault is that?) that they have to rewrite their drivers for the next major kernel revision.

Re:A Lesson In How To Use Market Forces

[JoeWalsh]

Even in my story I spent too much time fiddling with that NetGear, all because I didn't feel like driving back out to Best Buy.

You're absolutely right about us needing to return stuff that doesn't work as advertised. I've personally resolved to start treating computer hardware like I would any other physical item I purchase - if it doesn't work right away, without problems, I'll return it for a refund.

I made that resolution a few years ago. When Windows 98 was first released, my wife and I bought a new Compaq Presario unit for her parents. We brought it to their house, set it up for them...and the danged thing crashed right away. No problem, it's just Windows, right? Well, it was crashing every three to five minutes!
Instead of taking it back right away, we spent our entire week off trying to fix that thing. We spent time with Compaq on the phone, did the system restore thing a few times, got a replacement HDD shipped to us overnight, and all that stuff. Nothing made it stop crashing. Finally, at the end of our week off, we brought it back to the store and got another one. That one worked perfectly right away, right out of the box.

That's when we decided we'd never do that again. If some piece of hardware doesn't work properly right out of the box, we return it. Either it's poorly engineered and requires special attention to set it up, or it's broken in some other way. Whatever the case, my time's too valuable to spend on this sort of garbage, so I don't bother any more.

I hate returning things, but I hate wasting my time even more. I'm glad you brought this issue up.

Linux-tested.com already certifies hardware, badly

[Jade E.]

Recently shopping for a KVM, I saw several proudly bearing the linux-tested.com logo. In fact, *all* of the KVM's at this particular store had the logo. No big surprise, they're KVM's, right? I bought a Belkin 4-port OmniCube (Model F1D094) and thought nothing more of it.

That is, until I saw this story. I decided to go see what linux-tested.com had to say about this KVM (which works perfectly, btw.) The following are excerpts from their review:

The OmniCube has keyboard control of switching via a short sequence of keys followed by an onscreen display which has no effect upon the display as generated by the host CPU.

...

The OmniCube isn't packaged with software but does have internal key commands and on-screen display capabilities allowing for rapid switching of CPUs from the keyboard.

These, and several other references to the same feature, seem perfectly normal. Except for one thing: This model has no on-screen display! I've sat here and switched through my boxes every single way the switch supports (button on the switch and 3 different keyboard shortcuts), and there's absolutely no on-screen indication of any kind. Just in case mine is broken, I got out the box, but there's no mention of an OSD. (The models that do have OSD's have '-OSD' at the end of their model number, and aren't called OmniCubes.)

The model number on the KVM, and it's box, are the same one they gave in the review. The product name, OmniCube, is the same. So, question is, how could you possibly review a feature that didn't exist? Do they even look at the hardware they certify? I don't think I'd trust the 'linux-tested' logo for anything more complicated than a printer cable, cinsidering this obvious discrepancy.

The point (I knew I had one) is, we need to come down on certification places that don't use valid (and repeatable) testing procedures just as hard as we do manufacturers that claim linux support but don't follow through. I, personally, am going to be complaining to linux-tested.com and belkin both, and I think next time I go to buy hardware, given the choice between two similar pieces, I'll take one without a fake linux-tested logo over one from a company who thinks, 'Hey, let's make some money off this linux thingy by [selling|buying] tested logos for hardware!'

-Jade E.
I don't usually rant this much, just haven't had enough caffeine lately.

Other Linux nics

[underwhelm]

Try netgear FA310TX. It's a 10/100 nic that is tested on linux, comes with a tweaked tulip driver, instructions for linux installation and my personal endorsement.

Netgear did it right.

Re:Other Linux nics

[seebs]

There's no way I'll ever buy from Netgear again. Why? Because the FA-310TX is the model number of their DEC 21140 card. It is blatantly and seriously dishonest of them to continue selling a card with the same model number, which is not compatible with the old card. A driver that works for the old card may not work with the new card.

Can it be gotten to work? Certainly. At least, the NetBSD "tlp" driver should work with it. But I don't care; they owe me a new model number for their new model of card. :)

Support goes beyond "do they provide drivers". The question is, how much thought are they giving to people who are depending on their published information, or who are checking model numbers for compatability? In Netgear's case, it's "screw you, we put drivers on a disk, and that's the end of our responsibility."

I was using the D-Link cards based on the Via Rhine for a while. They're gone now, so I'm probably just gonna spend 2x as much on cards and use Intel EtherExpress cards; the current 10/100 models work in all 4 BSD's, presumably all the Linuxes out there, and probably everything else. They work well, they're cheaper than 3Com's overpriced crap, and the company is pretty good about adding revision #'s when they change hardware.

Re:Other Linux nics

[seebs]

Compaq may do it. 3Com has been pretty good about at least screen-printing revision numbers on boards, so you *can* tell that the 3C905 you got isn't the version listed as supported. Still, all of the vendors have some amount of this, the question is, how incompatible are they willing to be?

I would not have minded if NetGear had introduced the "FA-310TX Rev B".

A liccensed logo

[Felinoid]

Two thoughts on this....

One is you can have a trademarked logo that basicly means "We tested it and it works on Linux and we support Linux".
No testing agentcy needed...
This won't be as good as a central testing agentcy. However if the logo is enforced like the GPL has been enforced over the years then abusers of the logo can fix the problem by providing a software solution (a driver for the hardware) or fix the hardware. Eather way the problem is forgoten. This will get hardware support out a bit faster as well as companys will want to have the offical logo some will slap it on to discover it's a liccensed logo.

Second thought is.... consummer fraud....

Saying hardware works on Linux (or leaving you to believe it dose) by placing a logo to the effect on the pacage should be considered consummer fraud.

Re:There's no certification

[cwis]

Shouldn't be that hard to start a new web site that tracked Linux device support

Maybe you are looking for http://lhd.datapower.com/?

Quick query....

[Robert S Gormley]

...

all those vendors, and the other distributions... certifying a product that works on all of them.... think the attached software, more than the hardware... with all the different standards... /opt, /usr/local, etc etc, not to mention rc.d stuff... this is going to be a nightmare. Why do you think a lot of commercial software aims at RedHat? Because they have a reasonable semblance of guessing where to put things....

Is it right? No. But it's all that can realistically be expected.

How do you determine which distributions get 'power' in the venture? Does RobertGormleyLinux 1.1 earn me a spot?

Re:Quick query....

[Azog]

Good questions... Suggestions:

If RobertGormleyLinux can pony up 4 thousand dollars, it gets a spot. Or, if it is a non-profit, and has more than 4 thousand users, it gets a spot. Maybe not perfect, but it's a start, isn't it?

As far as difficulty of certification - yeah, it wouldn't be easy. That's why the certification group would need to have some money.

"Linux Compatible" would have to be defined rather carefully. There might have to be quite a bit of fine print. But...

At least when you went out to buy a bunch of 100-Mbit network cards, if there was a "Linux Compatible" penguin logo on it, and a note like

• This product is supported and tested on Linux Kernel 2.2.10 and up, as included in RedHat 5.2 or higher, SUSE 6.1 or higher, Mandrake 6.0 or higher, and Debian 2.5 or higher. Other distributions will likely work as well.
  

Frankly, that would give me a warm fuzzy feeling and I'd be a lot happier about buying those cards. Even if I was running RobertGormleyLinux 1.1 - after all, if I was eleet enough to run that, I would know what kernel version and patches I had... :-)

Obviously things are trickier for software. Still, I think it would be better than what the Linux world has now...

The rc.d stuff isn't that bad. There's two main styles of doing it (Sys V or BSD) and a few variations. Perhaps one of the things the hypothetical org could do is help standardize things, or write some nice flexible installation scripts that deal with the differences out there.

It couldn't hurt, could it?


Torrey Hoffman (Azog)

Maybe I'm pedantic...

[Robert S Gormley]

... but " Tested with Linux" does not mean "We provide Technical Support for Linux"...

Re: Professional Certification for Linux

[ai731]

I think it would be very beneficial for linux in general if some of the large, more trusted names in Linux got together to organize a Linux Certification system. The whole linux name would gain some credibility if Red Hat, Corel, Mandrake, Suse, VA Linux, etc., formed something by which they would have the power to give and take "Linux Certified" stickers or something and give out "Linux Certified System Administrator" certifications.

Linux Professional Institute is just starting to do this. They've written and beta tested exams for Linux System Administration 101 and 102. From the sample questions on the website, it looks like they've done a competant job. I'm planning to sit the exam in a month or two.

Certification needs to be judged by peer review. If competant people agree that the LPI only certifies competant people, then we have a de-facto standard that is worth something. It doesn't matter who set it up in the first place.

ai731

--

Re:There should be Certified Distro X hardware

[LenZ]

That's why I think there should be Distro certified hardware instead. (for example, SuSE supported would have a chameleon)

JFYI: SuSE already does this for quite a while (about a year). See our Hardware - Certification pages for more info about this. I can assure you - we bang those boxes hard, it is not just a simple test installation.

Re:Market is too broken - no central linux regulat

[Alik]

What about Linus? He's got the trademark on the word "Linux", IIRC. Would he not have the power to demand that a company making Linux-broken objects stop using his trademark (since it dilutes the mark to have it associated with stuff that doesn't work)? If this is not strong enough, it seems that he could simply trademark "Certified Linux Compatible" with a little Tux logo and tell us to look for that.

This is simple

[FascDot Killed My Pr]

Step 1: Does the box have just the Tux logo (goto 3) or does it also say something like "it works with Linux"? (goto 2)

Step 2: Tell tech support that the box says "it works with Linux" but it doesn't. Keep them on the phone until they fix your problem. If they refuse, tell the store manager where you bought the product, write a letter to the BBB and (depending on the situation) contact a lawyer.

Step 3) If it's just a logo, WRITE a PAPER letter to the company saying that you are considering buying one of their products, but you want to know what the penguin logo signifies. If they claim (in writing on paper) that it means "it works with Linux", goto step 2. If they don't respond or say "it's just decoration", tell the store manager where you bought the product and write a letter to the BBB. (this is assuming it really is Tux and not just a generic penguin--if it's generic you can't claim you were defrauded).

I don't know why people keep insisting on reinventing the wheel for these problems. This is not a Linux-specific issue. What if you bought a powerstrip and it pictured a man with a toaster and a laser printer plugged into it but the instructions said "never plug a toaster and laser printer into the same strip"? Same damn issue: misleading (if not outright false) claims made on a box.
--

Re:There's no certification

[Devil Ducky]

I know I don't have any fancy links to prove my points but here they are anyway.

RMS did NOT invent Free Software. In the beginning all software was free/open-source. The original hackers didn't care about closing everything up to keep it "secure" (Many still don't) That was the invention of the PHBs that followed. RMS crusaded to re-free the software.

That "side effect" of free software being changable is one of the original reasons for leaving the source open, and the reason the PHBs shut the source down (it's very hard to sell software that's open source.) The freedom of choice that RMS so vehemotly supports is the real "side effect."

In fact some of those old free programs are still in use today, just another side effect.

Devil Ducky

Use a Linksys

[nnet]

I use a Linksys KVM with *bsd, NT, and Linux. Works fine, and the box it comes in says Linux is supported, as does most, if not all of their products boxes, say. I haven't had a need to call their techsup though. :)

Re:Use a Linksys

[PD]

I'll second that. My four port Linksys switch runs very well.

I can't get it to switch with the keyboard though. I have to use the buttons on the front of the box. Their website wasn't any help to me either. I've never lost either a mouse or a keyboard in about 4 months of use.

Re:There's no certification

[Devil Ducky]

>Then why are you using c++ style commenting?

I use // instead of /* one line of text */
simply because I'm lazy, and GCC didn't complain.

That's why I code after all, too lazy to do anything else, and GCC didn't complain. :-)

Devil Ducky

Consumer power

[Digital_Quartz]

What can we do? The one thing consumers have always been able to do. Take it back, and demand your money back. If they won't give you your money, scream a lot. :)

Re:There's no certification

[Devil Ducky]

According to the story the device had a Tux Logo on it, it never said "Linux Compatible." And even if it did say that all they would have to prove is they had one succesful test on a linux machine and then they broke no false advertising laws only stretched a few.

The only obvious law they would have broken would be trademark infringement (Linux owns "Linux" and all of it's pronounciations :) But it is that infringement that is the threat behind the stance Microsoft, Novell, IBM, Apple, and everyone else who support 'Compatible' third-party products.

I don't think Linus would have the time, but he could nominate someone to look into these situations on his behalf. That someone would be helped by the entire community pointing out the offenders and then (s)he would test the complaint and take the neccessary action (a letter to the manufacturer should generally be enough.)

Devil Ducky

There IS a differnt Cert

[TheCarp]

I dunno if anyone else has mentioned it (looked didn't see it) but I believe there is has an "Open Hardware Certification" which does NOT garauntee Linux Compatibility but is actually better.

To get the certification vendors must publish enough information to write a driver and make it available to the public.

http://www.openhardware.org/

-Steve

But....

[NetJunkie]

As noted in my article submission, IO Gear couldn't give me a tested configuration. If they had said RedHat 6.2 with default kernel, etc..etc...that would have been fine. But they couldn't do that.

At least that shows they tested it...

[NetJunkie]

If it says it works with RedHat at least you could assume it would work with a comparable configuration on another distribution. My main problem with this company was that they couldn't even give me one configuration they tested it on.

Forget Test-based Certification

[hanway]

Don't bother with expensive test-based Linux certification. That's not going to work except for very few configurations (i.e. out-of-the-box Red Hat). Instead, Linux needs a certification that says all drivers are open sourced and all specs published, including support for all advertised features.

KVM switches & mice

[SoftwareJanitor]

We use Belkin OmniCube 2-port KVM switches here. Several people who use Microsoft Intellimice have had problems with the mouse not responding after a switch, requiring flipping back and forth before the mouse will respond again. I use a Logitech MouseMan, and I've never had such problems. I believe it is the mouse that is causing the problems, not the switchbox because I can switch the box with someone else's and it still works for me and occasionally goofs up for them.

Linuxcare Labs provides vendor nuetral testing

[sled]

<Disclaimer>
I work for LinuxCare.
</Disclaimer>

This service is currently being offered by Linuxcare Labs. We currently offer vendor nuetral product certification designed to demonstrate compatibility with the Linux kernel and other major subsystems of a GNU/Linux operating system environment. Working in this capacity I have learned about many of the challenges that come with trying to provide independent validation of Open Source based product.

There are many challenging questions to answer when certifying Linux/Open Source based products. For example, which distribution are tested against by default? How do you treat hardware that is only partially supported, i.e. 3D video acceleration, USB, fire wire, etc. How do you make a hardware vendor understand that the certification of their products depends on external factors over which they have no control, i.e. distribution packaging practices or the ability or willingness of Free Software developers to write a driver? Do you require everything to work "out of the box" or do you allow post installation configuration steps to be taken? For example, many sound cards on the market today won't work after a default installation of most distributions, and require that you download, compile, and install the latest version of ALSA to support the card.

Answering these questions is a constant balancing act between meeting the needs of the product vendor and delivering a true benefit to the consumer. In the end, certification loses its value if strict standards are not adhered to. However, at this point in the game it is difficult to convince a vendor to even consider investing in having their products tested under Linux without making it a very attractive proposition for them. What this usually translates to is going the extra mile to "make" a product work. When Linux compatibility testing is no longer optional for computer product vendors, the burden of finding out and documenting how to support a particular product will be shifted to the product vendor.