The Finnish company behind the open-source file system NTFS-3G, Tuxera, has announced it has signed a confidential intellectual-property agreement with Microsoft about creating a Linux exFAT driver which they will sell to Linux OEMs. They are in talks with Microsoft about an open source exFAT driver.
“The confidential Intellectual Property Agreement is basically about patents and giving us access to some Windows source code,” Tuxera chief executive Mikko V~A¤lim~A¤ki told ZDNet UK, “ExFAT is part of the forthcoming SDCX standard for flash cards, and we’ll be selling our driver to OEMs for devices like cameras.”
Tuxera is also talking with Microsoft about the possibility of an open source exFAT driver for Linux. “We’re talking to Microsoft about an open-source exFAT driver, but that’s not covered by the agreement,” V~A¤lim~A¤ki said, “We cannot sell end-user proprietary drivers, we can only sell exFAT on Linux to OEMs at present.”
When it comes to NTFS, Tuxera has a dual-licensing scheme, producing an open source variant and a proprietary one. The proprietary one has additional features making it a better fit for certain specific purposes. V~A¤lim~A¤ki claims there are no patent issues with NTFS. “Microsoft has never publically said anything about patent issues with NTFS. Our open-source NTFS driver has been available for 10 years, and our commercial driver for two.”
He added: “We’ll be licensing our Linux NTFS under the GPL, and we have an agreement with Microsoft. If you’re a user, you don’t need to worry about Microsoft. We’ll deal with them directly.”
This is wonderful news. Now… I wonder what the catch will turn out to be. It’s not really a matter of ‘if” but of “what”. I cannot think of any alternative which would not also validate the existence of Santa Claus, the Tooth Fairy, and possibly the Easter Bunny.
Edited 2009-08-26 18:17 UTC
The catch is it’s not free:
“a Linux exFAT driver which they will sell to Linux OEMs“
Well, I was really thinking of the open driver that they are “talking to Microsoft” about.
Edited 2009-08-26 18:39 UTC
As I said in the other post:
Open doesn’t always mean free.
Besides, this driver might not even materialise given they’re still only at the “talking to MS” stage – let alone proof that any such driver would be GPL
Edited 2009-08-27 15:31 UTC
From the article:
“We’re talking to Microsoft about an open-source exFAT driver, but that’s not covered by the agreement.”
Open-source does pretty much mean Free. Although, as you note, it could be under CDDL or some such, which would be a show stopper for a kernel module. Though if it is a FUSE filesystem, any OSS license would do. And of course, it is at this point only talk. But that is still encouraging, and about as wonderful a news as anyone could expect at this point.
But I did start out wondering what the “catch” was. And there is certainly plenty of room for them at this point. But I’m in a better mood today and trying not to be overly-negative.
Actually, it doesn’t even need to be an OSS license if you are using FUSE. So basically, the GPL won’t “block” Linux from accessing any filesystem for which FUSE driver can be written.
It does if we want an open-source driver for Linux. You are, of course, correct that even a proprietary driver using FUSE would be legal to distribute. But it wouldn’t satisfy many FOSS folks. My position is that any driver which doesn’t have a cloud of uncertainly hanging over it is better than one that does. But I would *much* prefer to have one under some sort of open license, copyleft or no.
Directly from the summary:
“They are in talks with Microsoft about an open source exFAT driver.”
It may be free, lets wait and see
’cause if it’s not, it won’t be bought
We’ll trim the FAT, we won’t use that
And MS won’t care, won’t notice a hair.
Poem by Jake.
Open Source doesn’t always mean free though.
The fact is, there is no evidence (other best wishes / high hopes) that any such driver will be GPL.
if it is opensource, it has to be free. If the source is available under a BSD or LGPL style license, then somebody can then re-release it as free.
Therefore, if it is opensource, it will be free, eventually.
BSD and LGPL are only the tip of the iceberg of open source licences.
There’s some that aren’t free (eg MS have released code under their own open source licence that’s only available for a fee and isn’t redistributable)
Then that isn’t an opensource license. Also, I said
.
That should cover most of the OSS licenses that they are likely to use.
Edited 2009-08-27 16:45 UTC
That depends on your definition of “open source”
Sometimes, open source can just mean that the source if available but not licensed for distribution.
I know you did, but you were making assumptions.
In reality they could use which ever licence them and Microsoft agree upon.
If they are in talks with MS about an OPENSOURCE driver, then that puts restrictions on what kind of license that they are talking about.
They can’t charge for the source if it is an OPENSOURCE license. They can’t restrict redistribution of the code if it is an OPENSOURCE license.
That is what an OPENSOURCE license means, regardless of the particular flavour of license.
Don’t assume MS will restrict the source, or charge for it, wait until those talks are finished, and then we’ll see. The talks will either be successful, and it will be opensourced, or they won’t be, and it won’t.
All open source means is that teh source is openly available.
It doesn’t mean that the product is free to distribute for free.
Many Linux/BSD geeks get so carried away with their own particular blends of open source licences that they forget that you can charge for a product and offer the source code (many Unix’s (used to) work on this model if fact)
Edited 2009-08-31 23:48 UTC
It’s not wonderful news and the catch is that it appears to legitimise more and more the need for you to go to a third-party, in this case Microsoft, and ask for permission to develop something standalone for the purposes of interoperability. Microsoft doesn’t want a freely available and unencumbered ExFAT implementation running wild and I doubt whether they’ll cooperate over creating a driver that can be in the kernel. This should never be necessary for open source software.
Quite frankly, if these people ever make contributions to the Linux kernel or open source software in any way then they cannot be trusted because they’ve been tainted. It probably won’t matter too much in the end though because someone else will.
Edited 2009-08-26 20:35 UTC
NTFs-3G is a FUSE driver (IIRC), and exFAT can be too. So, this is basically pretty harmless stuff. If some OEM insists on using exFAT, I won’t complain when they have to pay up for licenses. In any case, it wont taint the kernel.
FUSE drivers are not exactly the best way of ensuring that you can handle a particular filesystem. The worrying thing about this is that when someone else proposes an in-kernel driver, as will inevitably happen, then this ‘deal’ will be used as a rod to beat that project as well.
Well, no it isn’t because that completely subverts the whole point of open source software – the ability to create freel developed, freely available and freely used implementations.
How do you expect a distribution to ‘pay up’ for licenses? Do you use a free distribution as many do? Whoops. Not any more.
Additionally, if and when most cards start using ExFAT then OEMs won’t exactly be ‘insisting’ on using it, which is a bizarre choice of word.
Edited 2009-08-26 23:49 UTC
Isn’t it a good thing? It keeps the possibly contaminated code out of the kernel.
Wasn’t that subverted when they standardized on this crap? If it’s either this or no support at all, I’d rather pick this.
I understood only OEMs need to pay. If distros had to pay, they would obviously not install this driver and you’d have to install it separately. Just like you can’t play all multimedia content out of the box right now.
They are part of the industry that advocated, or silently approved this standard. It seems reasonable to expect them to pay.
Having a legal Linux driver is *good* news. The fact they are even talking to MS about a free driver is the *wonderful* part. The laws are the laws. If what the laws say is wrong or bad, it is the laws that need to be changed. Obsessing upon what a driver for Linux legitimizes or doesn’t legitimize is not only a waste of time and energy, but qualifies as cutting off one’s nose to spite one’s face.
We had a discussion the other day about why Linux is making such slow progress on penetrating the desktop market, and I suggested that we, the small minority, keep sabotaging ourselves by demanding that everyone else do things exactly our way before we will play. And your attitude is an excellent concrete example of that problem.
Edited 2009-08-26 23:43 UTC
Legal according to whom? Are a lot of existing drivers somehow illegal?
Why do they need to talk to Microsoft, unless of course, Microsoft are going to allow code contributions to a freely developed implementation? I was always taught that relationships were two-way. There are names for ones that aren’t.
What laws would these be, exactly? You can’t just assume illegality because the very company you are pleased that they are talking to has cast aspersions.
Unless it’s good for open source software and the ability to develop and distribute free software then it’s not good for Linux. That’s the way it is and that’s what has contributed to Linux’s success. The bed has been made there. If that integrity is lost then you get an uncoordinated mess. How often has the binary blob method been successful for driver development and distribution over the years?
That’s one of the funniest things I’ve seen written about the subject because it’s the absolute polar opposite of what’s required. Take a look at Microsoft’s history themselves and ask yourself how many times they have willingly played by other peoples’ rules. Did they chuck the towel in over Office and start using WordPerfect and Lotus 1-2-3’s file formats because that’s the way the world was at the time? No. Did they simply reuse another LDAP implementation when coming up with Active Directory? No. They worked out what they were about and followed through. What was that about ‘our’ way again?
The first rule at being successful is to work out what you’re about and what makes you different, work from there and protect that difference. If you don’t you spend all your time being divided and conquered as you fanny about playing by someone elses’ rules. Desktop Linux has enough work ahead of it without thinking that that is somehow the mystical route to success, and even worse, getting completely sidetracked into thinking that that’s why things haven’t happened so far.
Edited 2009-08-27 00:27 UTC
exFAT is a crippled crap anyway. I always reformat my USB/SD drives with EXT2/UFS/FFS … and I don’t use any proprietary formats from cameras, etc
I’ve never seen a camera that uses exFAT, but I’m sure they’re coming and chances are Linux, woun’t be able to read its files. There was work to build a FOSS exFAT driver that had reached a point of being able to reliably READ (but not write) files as late 2008, but work on it seems to have stopped and the patchset will probably never make it into the mainline kernel. Given the fact that the developers of this new driver are “in bed” with Microsoft, I somehow have strong doubts as to whether any free exFAT driver will ever come from them. Given that these same people wrote ntfs-3g, I wouldn’t be surprised if the free version went away as well.
“The confidential Intellectual Property Agreement is basically about patents and giving us access to some Windows source code,” Tuxera chief executive Mikko V~A¤lim~A¤ki told ZDNet UK, “ExFAT is part of the forthcoming SDCX standard for flash cards, and we’ll be selling our driver to OEMs for devices like cameras.”
I always worry just a little bit about people embracing non-open standards.
I second that – I’d love to see an international agreement where any technology becomes an open standard or part of an open standard, people must be allowed to implement it without paying royalties. Its these ‘agreements’ that create walled gardens and stifle innovation and interoperability – entrench existing monopolies and create new ones.
Business is business…
Still, anything that isn’t open is *not* a standard, but an enforced monopoly. It does not deserve the name standard.
Couldn’t they just use BFS anyway ?
But why is the file system included in memory card standard? Does this mean that we cannot use any other file systems on those cards? I guess not. Then what is the fuzz about – any device can just stick to crappy old FAT – it supports file systems up to 2TB. Even if Windows wouldn’t let you format more than 32GB. I have formatted 200GB FAT partition under Mac OS X and it works on Windows.
So, what uses exFAT again? Most things I’ve seen just use the old fashioned FAT that everyone reads/writes.
ExFAT has become the default file system for SDXC; what I think is crappy is the double dipping that is involved. Take AAC for example; encoders pay, devices pay, software audio players pay. Why should I have to pay for AAC access on my media player only to have to pay another royalty so that I can encode and put music on there?
Same thing is going to happen; SDXC will pay a royalty to Microsoft, then the customer will pay when it comes to using it on a non-Microsoft operating system – and so the double dipping continues on indefinitely.
Edited 2009-08-27 03:23 UTC
>> Tuxera, the Finnish company behind open-source file system NTFS-3G (from the article)
NTFS-3G is a filesystem? I thought it was just a driver which allowed use of the NTFS filesystem.
but a *confidential*? Damn, FAT is not only a rocket science and well know, reverse-enineered, documented in many existing open source project etc… What othe confidential information can they have?