Linked by Thom Holwerda on Tue 18th Sep 2012 21:45 UTC
PDAs, Cellphones, Wireless Microsoft and RIM have announced that RIM has licensed Redmond's exFAT patents. The press release contains a ridiculous amount of hyperbole nonsense, and if you translate it into regular people speak, it basically comes down to RIM paying Microsoft protection money for stupid nonsensical software patents. Ridiculous articles like like this make it seem as if we're talking about patents on major technological breakthroughs, but don't be fooled: this is because for some inexplicable reason, we're using crappy FAT for SD cards.
Order by: Score:
Metadata
by pgquiles on Tue 18th Sep 2012 22:24 UTC
pgquiles
Member since:
2006-07-16

The reason we are using FAT for SD cards is FAT consumes little space in metadata and is supported by all the major operating systems (Windows and Mac).

Don't forget most people (and OEMs) only moved from FAT32 (which is essentially FAT, only consuming a few more blocks for long names) to NTFS when hard disks became larger than 32 GB. I expected a similar move for SD cards but exFAT, which is lighter than NTFS, may make that different.

Problem is the exFAT driver on Linux is very very slow (and affected by patents), unless you purchase the commercial implementation by Tuxera.

Edited 2012-09-18 22:25 UTC

Reply Score: 5

RE: Metadata
by _txf_ on Tue 18th Sep 2012 22:38 UTC in reply to "Metadata"
_txf_ Member since:
2008-03-17

Problem is the exFAT driver on Linux is very very slow (and affected by patents), unless you purchase the commercial implementation by Tuxera.


Not just exFat. FAT is protected by patents as well, I imagine all Android/Linux makers got nastygrams with fat being some of the infringing patents.

ExFAT is basically just regular FAT with support for larger volumes and file sizes plus some minor features cribbed from more modern filesystems.

Edited 2012-09-18 22:39 UTC

Reply Score: 2

RE: Metadata
by PieterGen on Tue 18th Sep 2012 22:42 UTC in reply to "Metadata"
PieterGen Member since:
2012-01-13

Most SD-cards never get exchanged between devices. So, why doesn't RIM simply choose a good an affordable (read: free) file system? ext3, etx4, reiser, jfs....

Why choose exFAT, which is a)worse and b)more expensive?

I mean, Microsoft isn't exactly known for state of the art filesystems, or did I miss something?

Reply Score: 1

RE[2]: Metadata
by pgquiles on Tue 18th Sep 2012 22:48 UTC in reply to "RE: Metadata"
pgquiles Member since:
2006-07-16

Because many people do extract their SD card from their camera or mobile phone and insert it into their laptops.

Try to do that with ext3 and you've got yourself into three problems:

- Not supported out-of-the box on Mac and Windows

- Even after installing a driver, the support is not that good. Unless you go for a commercial implementation, such as Paragon's, and that means royalties. And if you are going to pay royalties, well, wtf, pay them to Microsoft and get FAT, which does not neeed a driver!

- Some people will try the SD card on a computer where the driver is not installed, they will receive the infamous "unknown partition" message and bitch about having lost all their data

So essentially going for FAT or anything which is supported out of the box by Windows and works acceptably well is the best business decision. Props if it also works out of the box and reasonably well on Mac. Linux? No need to care about them.

Reply Score: 7

RE[3]: Metadata
by MollyC on Tue 18th Sep 2012 23:00 UTC in reply to "RE[2]: Metadata"
MollyC Member since:
2006-07-04

You wrote and submitted your post while I was posting my brief response to PieterGen . You said it better than I did. ;)

Reply Score: 5

RE[3]: Metadata
by Lobotomik on Wed 19th Sep 2012 05:33 UTC in reply to "RE[2]: Metadata"
Lobotomik Member since:
2006-01-03

How about one of the filing systems supported by OSX, like UFS? It is a powerful filing system, with source code available under a BSD license, and compatible out of the box with Macs. It would get along fine with the Linux and iOS kernels that make up 95% of the smartphone market, and a huge slice of the HD-Tv market, the set-top box market and the embedded market in general.

Only Microsoft would be left out, but they have already stated that the BSD license is palatable to them, and if this were official they might could keep face by adopting a "multimedia standard"

Reply Score: 1

RE[4]: Metadata
by lucas_maximus on Wed 19th Sep 2012 16:45 UTC in reply to "RE[3]: Metadata"
lucas_maximus Member since:
2009-08-18

Because FAT is already good enough. FAT32 is still fine for the majority of applications, exFAT can supports silly file sizes.

I suggest you read the unix haters handbook, the principle is called "worse is better".

I haven't worried about what filesystem I should choose since I last tried a Gentoo Install, or when Ext4 would forget to actually put the file onto the disk.

Edited 2012-09-19 16:47 UTC

Reply Score: 2

RE[4]: Metadata
by Alfman on Wed 19th Sep 2012 17:31 UTC in reply to "RE[3]: Metadata"
Alfman Member since:
2011-01-28

Lobotomik,

"How about one of the filing systems supported by OSX, like UFS? It is a powerful filing system, with source code available under a BSD license, and compatible out of the box with Macs. It would get along fine with the Linux and iOS kernels that make up 95% of the smartphone market, and a huge slice of the HD-Tv market, the set-top box market and the embedded market in general."

It's good that we're finally starting to think about alternatives to the FAT lock in. Someone else may be able to highlight whether UFS could be suitable or not, since I have no experience with it. However there is a potentially large problem with your logic: just because "linux" supports UFS doesn't mean it's ok to extrapolate that 95% of the smartphone market, embedded market, etc can already support it. These would likely need to be re-flashed to get the driver, they might not have sufficient resources for a new driver. Even if the stock kernel did include UFS, there still could be issues with the user-space tools on these embededed devices. My own linux desktop distro doesn't even have ufsutils installed (mkfs.ufs, fsck.ufs, etc). It's doubtful that any UFS combination has ever been tested on most embedded devices.

Not to discourage progress, but I think switching file systems is going to be to be an interoperability mess. The safest way to transition would be to start explicitly supporting another file system today and keeping FAT as the default. In a few years time when the alternative FS is widely supported, then it should become the default. But how do we get all manufacturers on board with this plan? It almost takes a monopoly to get the ball moving.

Reply Score: 3

RE[3]: Metadata
by JAlexoid on Wed 19th Sep 2012 10:09 UTC in reply to "RE[2]: Metadata"
JAlexoid Member since:
2009-05-19

Actually Microsoft "was selected" by the SD Assiciation to make the exFat as the mandatory standard for SDXC.

Reply Score: 4

RE[4]: Metadata
by dsmogor on Wed 19th Sep 2012 12:49 UTC in reply to "RE[3]: Metadata"
dsmogor Member since:
2005-09-01

Hey, if that's part of a standard shouldn't patents in question be licensed on RAND terms?

Reply Score: 5

RE[5]: Metadata
by kurkosdr on Thu 20th Sep 2012 11:21 UTC in reply to "RE[4]: Metadata"
kurkosdr Member since:
2011-04-11

"Hey, if that's part of a standard shouldn't patents in question be licensed on RAND terms?"

They are. Much like the H.264 patents are.

Reply Score: 2

RE[2]: Metadata
by MollyC on Tue 18th Sep 2012 22:55 UTC in reply to "RE: Metadata"
MollyC Member since:
2006-07-04

RIM likely did a cost-benefit analysis and found that paying a (very cheap, BTW) license fee for exFAT was the best option for them.

Reply Score: 4

RE: Metadata
by UltraZelda64 on Wed 19th Sep 2012 01:44 UTC in reply to "Metadata"
UltraZelda64 Member since:
2006-12-05

Problem is, FAT gets incredibly inefficient the larger you make your partition. You would have to be nuts to actually want to format a 128GB drive or partition with FAT, and I wouldn't really want to make a FAT partition with a size of 64GB either. There's a reason Microsoft has placed an artificial limitation in Windows and FDISK versions starting with Windows 98 (though maybe excluding Windows ME): FAT SUCKS on larger volumes.

It's not just to get you over to NTFS (although that is no doubt the primary reason). But in this case, they're actually right: it does make more sense to use NTFS instead of FAT on larger partitions. Far more sense, in fact. Cluster sizes jump from 16K to 32K in volumes over 32GB, leading to more slack space, especially if you have a decent number of small files. The larger the partition is, the larger the file allocation table becomes, to the point of being a massive waste of space on its own. The FAT, quite literally, gets fat.

Only potential exception: flash memory devices. NTFS is not designed to work well with those and their finite number of writes per block. I guess if you're only going to put a bunch of large videos and similar files on it you might be fine, but watch out for another potential problem: FAT has a file size limit of 4GB. Sorry, no full-size DVD images here.

http://www.pcguide.com/ref/hdd/file/partSlack-c.html [slack space]
http://www.pcguide.com/ref/hdd/file/partCluster-c.html [partition/cluster sizes]
http://www.pcguide.com/ref/hdd/file/partFAT32-c.html [FAT sizes]
http://www.pcguide.com/ref/hdd/file/part.htm [the whole site]
http://www.allensmith.net/Storage/HDDlimit/FAT32.htm [another site with some information]

I was forced at one point to use FAT32 to format one of my external hard drives to be able to use it to play music from while playing my Xbox 360. If I had the choice to use another file system, I would not have resisted at all. Even NTFS would have been nice--another Microsoft file system that is much faster, more efficient and more reliable with large drives, but the system would only accept FAT.

FAT just sucks, and exFAT is every bit as bad when it comes to these patents that Microsoft was granted and is enforcing. Not to mention its support outside of Windows before Vista sucks, as does its use as a general cross-platform-compatible "universal" file system.

Edited 2012-09-19 01:58 UTC

Reply Score: 4

RE[2]: Metadata
by Lobotomik on Wed 19th Sep 2012 05:19 UTC in reply to "RE: Metadata"
Lobotomik Member since:
2006-01-03

Not that I think that FAT or exFAT is anything other than a POS, but who cares about cluster sizes of 32K when the files are mostly going to be media files a couple megs long (pics or songs)?

Reply Score: 3

RE[3]: Metadata
by UltraZelda64 on Wed 19th Sep 2012 06:24 UTC in reply to "RE[2]: Metadata"
UltraZelda64 Member since:
2006-12-05

I don't know about you, but my Android device doesn't even take pictures that are "a couple megs long." ;)

They're far smaller and more compressed than that. I'm surprised to see it produce one that is even 1MB in size. I don't know about you, but I also have other small file like simple text files that I make and use and on a regular basis on my phone. Surely there might be some small PDF files, and many configuration files scattered around too. I tried to format the microSD card with ext2 right when I got the phone and ditch the pre-formatted FAT file system, but Android won't recognize it and just asked to re-format it with FAT. That should not be the case given that it runs the Linux kernel; hell, version 2.2.2 of Android that it runs uses ext4 as the OS file system. So, well, I'm forced into FAT. Again. Yay.

I would like to ditch FAT on every machine I have, except in a rare FreeDOS virtual machine for nostalgia and fun, but it won't happen when companies keep supporting this wretched ancient family of file systems. And then you're got Microsoft extending the creaky thing far beyond its usefulness. Having had to use FAT32 with Win9x and experience (or should I say, put up with) the sudden disappearance of random system files and other files on a regular basis, I definitely have something against FAT. It is, IMO, complete trash. It needs to be accepted as what it is: obsolete technology that has long outlived its welcome.

There is no reason for a Linux-based device to enforce the use of such an old non-native DOS file system, and it's crazy that a piece of Microsoft hardware running a modified Windows NT kernel (the Xbox 360) requires FAT32 and will not operate with NTFS.

Edited 2012-09-19 06:38 UTC

Reply Score: 1

RE[4]: Metadata
by moondevil on Wed 19th Sep 2012 12:17 UTC in reply to "RE[3]: Metadata"
moondevil Member since:
2005-07-08

There is no reason for a Linux-based device to enforce the use of such an old non-native DOS file system, and it's crazy that a piece of Microsoft hardware running a modified Windows NT kernel (the Xbox 360) requires FAT32 and will not operate with NTFS.


Interoperability?

Which other filesystem do you know that is so universally accepted by most devices as FAT and its successors?

Reply Score: 4

RE[5]: Metadata
by UltraZelda64 on Thu 20th Sep 2012 00:21 UTC in reply to "RE[4]: Metadata"
UltraZelda64 Member since:
2006-12-05

That is NO EXCUSE for an operating system based on the Linux kernel to not just read and use an ext2/3/4-formatted microSD card that I formatted in Linux myself. Similarly, it's no excuse for a system with a modified NT-based kernel to refuse to use an NTFS file system that I formatted in Windows myself.

NTFS "interoperability" with systems other than Windows has been pretty decent for a while now thanks to FUSE and NTFS-3G with full read/write capability, although Microsoft would prefer that it wasn't. And even if it wasn't, considering Microsoft would like you to believe that no operating system other than their own exists in the world--even from that perspective it makes absolutely NO SENSE why FAT is required of a modern video game system released by them in 2005.

Consider all of the technical and graphical specifications the Xbox 360 has... some impressive stuff... and for external storage devices, it supports... FAT32?! WTF?!? Even Windows has moved away from it--I haven't used FAT partitions on my hard drives starting with Windows XP over a decade ago, and being the default I doubt many other people have either. The immensely improved reliability was well worth it (no more randomly lost files and fewer failed Windows boots), but the performance is much better too (as long as you stay on top of the excessive fragmentation and resulting slowdown typical of Windows...).

To put it simply, unless you are using DOS on ANY of your machines, you really do NOT need FAT for interoperability with all of your computers. Are you? I'm not. It's 2012, not 1990. And if a device (cell phone, digital camera) runs a kernel that natively supports its own file system, you should NOT be forbidden from formatting your storage device to that file system and using that instead of the lowest common denominator (FAT), especially in a time when dozens or hundreds of gigs have long been typical and FAT has been losing relevance for years. Hell, when I left Windows in late 2006 I was using NTFS, ISO9660 and the occasional UDF... no sign of FAT, anywhere.

Either way, my point still stands. If no one steps up and just says "fuck FAT" and uses something else instead, FAT will continue to be used indefinitely. It's cheap, it's simple, it's unreliable and inefficient and it's crap, but the first two parts seem to make up for the rest for most companies. There is no need for it in many cases, and once people start using (or even allowing) other file systems on their devices FAT will finally fade away. If not, we'll be constantly tormented by it on virtually all portable devices for decades to come, despite many better alternatives being available.

Of course, it doesn't help that Microsoft doesn't support any disk file systems other than their own three... other people have already brought this up as being anti-competitive, and I agree. Would it kill them to implement read/write support for ext2/ext3/ext4/UFS? Microsoft wants to keep everyone on FAT for their portable storage, so once flash drives are commonly seen in sizes like 128GB, 256GB, 512GB and bigger and capacities of 64GB and lower become more uncommon, they will already have the companies by the balls and easily get them to start using exFAT by default.

Then after that comes hardware support--and by hardware support, I mean exclusive exFAT support, because at such capacities any other version of FAT would make zero sense. And with the active patents on the exFAT file system... well, you can guess what will happen next if someone implements it and doesn't pay Microsoft the toll. By then, it'll be time for Microsoft to start cashing in the big bucks, whether a company pays Microsoft for a license or they end up being taken to court (assuming the court doesn't invalidate the patent(s) in the process).

Reply Score: 2

RE[6]: Metadata
by moondevil on Thu 20th Sep 2012 06:51 UTC in reply to "RE[5]: Metadata"
moondevil Member since:
2005-07-08

You know that there are many other types of computers besides desktops, right?

Reply Score: 3

RE[7]: Metadata
by UltraZelda64 on Thu 20th Sep 2012 08:19 UTC in reply to "RE[6]: Metadata"
UltraZelda64 Member since:
2006-12-05

You know that there are many other types of computers besides desktops, right?

Yes. And you know there are other file systems besides FAT, right?

Coding in support for a few file systems more modern than that creaky old thing developed back in the 1970s with 8.3 file names and a kludge for what has become known as "long filenames" wouldn't kill anyone, would it?

Is it really so much to ask that the portable devices we actually spend money on be designed to accept the small handful of file systems that are native to the operating systems installed on our desktops/laptops/tablets/etc., which they are supposedly designed to connect with in the first place? Or even the file system(s) native to the kernel the fucking device itself it running (Android=ext2/3/4)? If it's removable storage (SD, microSD, etc.), you should be given the choice. Simple as that.

It's not like they would have to support microwave ovens, traffic light control systems, space shuttles, and supercomputers. Just the types of machines that normal people would plug a typical cell phone, camera or other portable device into using a USB cable.

Just to make it clear: Supporting other operating systems does NOT fucking mean you have to immediately drop FAT support, and therefore all of those Windows machines already out there that Microsoft continues to stubbornly and anti-competitively not allow interoperability with the rest of the computing world.

Reply Score: 1

RE[4]: Metadata
by lucas_maximus on Wed 19th Sep 2012 12:43 UTC in reply to "RE[3]: Metadata"
lucas_maximus Member since:
2009-08-18

Xbox 360 does not use a Modified Windows NT kernel.

According to Microsoft, it is a common misconception that the Xbox and Xbox 360 use a modified Windows 2000 kernel.[36] They claim that the Xbox operating system was built from scratch but implements a subset of Windows APIs. The idea that it does, indeed, run a modified copy of the Windows kernel still persists in the community


http://en.wikipedia.org/wiki/Windows_NT

Be glad to be able to use FAT 32 because I had to use older filesystems ... Sun Ray Machines don't even recognize FAT32, they only recognise FAT12 and FAT16.

Edited 2012-09-19 12:53 UTC

Reply Score: 2

RE[5]: Metadata
by UltraZelda64 on Thu 20th Sep 2012 00:25 UTC in reply to "RE[4]: Metadata"
UltraZelda64 Member since:
2006-12-05

If the Xbox 360 does in fact *not* actually use a modified version of the NT kernel, that is still no excuse for using FAT when NTFS was already well on its way to becoming the PC standard if it wasn't already--and default (due to Windows XP)--well before the time of the system's release.

I briefly read the page linked from the Wikipedia article you pointed out, and honestly... the way it was written (vaguely), I'm not so sure it's talking about the *kernel* as much as it's referring to the complete *operating system* that the machine runs. I would have to guess it's talking about the whole deal. Duh, obviously it doesn't run a complete Windows OS--I wasn't implying that at all. But I seriously doubt that Microsoft built a brand-new kernel 100% from scratch, and surely they didn't take much if anything from the DOS kernel when they've got NT and would be better off rewriting parts that are not already in it.

That blog entry doesn't really make it clear (to be fair, I just kind of skimmed through it), but it sounds to me like they're talking about the system as a whole... I see no mention specifically of the kernel, which I still would assume is NT-based. It does mention that the OS was built from the "ground up," but if I wanted to bad enough I could build an OS from the ground up based on the Linux kernel. It might share the userland and have the same kernel as the rest, but hey... it'd be built from the ground up.

Reply Score: 0

RE[6]: Metadata
by lucas_maximus on Thu 20th Sep 2012 09:29 UTC in reply to "RE[5]: Metadata"
lucas_maximus Member since:
2009-08-18

You said it ran a cut down version of NT or something along those lines, I can't be arsed quoting from up the comment pyramid.

I just corrected you.

The target demographic of the Xbox 360 don't really care about the underlying file-system. FAT is good enough for playing some MP3 and some Movies.

Reply Score: 3

RE[7]: Metadata
by UltraZelda64 on Fri 21st Sep 2012 00:17 UTC in reply to "RE[6]: Metadata"
UltraZelda64 Member since:
2006-12-05

Thanks for the "correction" if it really is one, but like I said, the blog entry that Wikipedia references doesn't seem to prove my original point. I believe my original quote that you were referring to was:

There is no reason for a Linux-based device to enforce the use of such an old non-native DOS file system, and it's crazy that a piece of Microsoft hardware running a modified Windows NT kernel (the Xbox 360) requires FAT32 and will not operate with NTFS.

As you pointed out, the Wikipedia article you linked to claims that the Xbox 360 does not use a modified NT kernel, but I'm not so sure that I believe it without proof to back it up. The "proof" the Wikipedia article has is the following reference in the form of a blog entry on MSDN, which is what I was referring to in my previous reply:

http://blogs.msdn.com/b/xboxteam/archive/2006/02/17/534421.aspx

That's inconclusive at best. I don't see anything specifically about the kernel itself and its design. I like Wikipedia, but this just seems like one of its weak points, and this seems to be a good example of it.

Edited 2012-09-21 00:18 UTC

Reply Score: 1

Where is the EU?
by Bill Shooter of Bul on Tue 18th Sep 2012 22:39 UTC
Bill Shooter of Bul
Member since:
2006-07-14

I don't understand why the EU hasn't raised a stink about this anticompetitive behaviour. We are using stinky FAT on our sd cards because MS scammed its way into a monopoly and then F'd everyone else over to get its technological underpinnings adopted as necessary standards for interoperability.

I'd love for the US DOJ to get involved again as well, but that's kind of like waiting for Godot.

Reply Score: 4

RE: Where is the EU?
by _txf_ on Tue 18th Sep 2012 22:42 UTC in reply to "Where is the EU? "
_txf_ Member since:
2008-03-17

We are using stinky FAT on our sd cards because MS scammed its way into a monopoly and then F'd everyone else over to get its technological underpinnings adopted as necessary standards for interoperability.


I seem to remember that Rambus got its ass handed to it for that very same behaviour. They submitted tech to JEDEC as a standard, which everybody then implemented. Rambus subsequently sued everybody, initally won, but then got slapped for anticompetitive behaviour.

Edited 2012-09-18 22:43 UTC

Reply Score: 1

RE[2]: Where is the EU?
by pgquiles on Tue 18th Sep 2012 22:49 UTC in reply to "RE: Where is the EU? "
pgquiles Member since:
2006-07-16

When did Microsoft submit FAT as a standard for anything to any standardizing group?

Yeah, never, so that's the difference between MS and Rambus.

Reply Score: 4

RE[3]: Where is the EU?
by saso on Tue 18th Sep 2012 23:09 UTC in reply to "RE[2]: Where is the EU? "
saso Member since:
2007-04-18

When did Microsoft submit FAT as a standard for anything to any standardizing group?

Yeah, never, so that's the difference between MS and Rambus.


True, the comparison to Rambus is inadequate, however we do see MS abusing their dominant position in the PC market here. In order to allow for interoperability, you need to implement their proprietary filesystem, which is patent encumbered. Not doing so puts you at a significant disadvantage in the market. Not saying that this isn't exactly established practice in the industry though (such as standards being encumbered by FRAND nonsense patents and the like...)

Reply Score: 2

RE[4]: Where is the EU?
by pgquiles on Wed 19th Sep 2012 05:29 UTC in reply to "RE[3]: Where is the EU? "
pgquiles Member since:
2006-07-16

we do see MS abusing their dominant position in the PC market here. In order to allow for interoperability, you need to implement their proprietary filesystem, which is patent encumbered.

Paying royalties in order to have interoperability is perfectly fine and the EU will not do anything about that.

The EU would only intervene if Microsoft wouldn't want to license exFAT to others, or if the royalties were way too high.

Remember, this is the same EU which has been trying to approve software patents for years.

Reply Score: 1

RE[5]: Where is the EU?
by saso on Wed 19th Sep 2012 10:41 UTC in reply to "RE[4]: Where is the EU? "
saso Member since:
2007-04-18

Paying royalties in order to have interoperability is perfectly fine and the EU will not do anything about that.

I know and I object to that. It stifles innovation in open-source which is forced to either pay up (using money they don't have) or remain forever unable to compete. I'd be calmer if only the inventions really were patent-worthy, but a patent on some primitive algorithm for 8.3 name handling is just ludicrous...

The EU would only intervene if Microsoft wouldn't want to license exFAT to others, or if the royalties were way too high.

I know, that's why I said "FRAND nonsense patents".

Remember, this is the same EU which has been trying to approve software patents for years.

That's the European Commission, not the "EU" and not the European Parliament. The EC is a body of appointed bureaucrats that aren't directly responsible to the electorate, so it's quite obvious they are the first go-to place for corrupting interests.

Reply Score: 3

RE[3]: Where is the EU?
by JAlexoid on Wed 19th Sep 2012 10:12 UTC in reply to "RE[2]: Where is the EU? "
JAlexoid Member since:
2009-05-19

Luckily for Microsoft SD Card Assiciation has no FRAND requirements.

Reply Score: 4

RE[4]: Where is the EU?
by dsmogor on Wed 19th Sep 2012 12:52 UTC in reply to "RE[3]: Where is the EU? "
dsmogor Member since:
2005-09-01

An that's is the crux of the problem.

Reply Score: 2

RE: Where is the EU?
by kurkosdr on Wed 19th Sep 2012 12:03 UTC in reply to "Where is the EU? "
kurkosdr Member since:
2011-04-11

I don't understand why the EU hasn't raised a stink about this anticompetitive behaviour. We are using stinky FAT on our sd cards because MS scammed its way into a monopoly and then F'd everyone else over to get its technological underpinnings adopted as necessary standards for interoperability."

Because MS licenses their FAT and exFAT patents under "FRAND" (for EU countries that have softpatents), much like MPEG LA licenses their stuff under "FRAND", so all is well.

Reply Score: 2

RE[2]: Where is the EU?
by Bill Shooter of Bul on Wed 19th Sep 2012 14:47 UTC in reply to "RE: Where is the EU? "
Bill Shooter of Bul Member since:
2006-07-14

Yes, I understand that more now than when I posted. Except that its really ill gotten gains. People need to use their technology because of their monopoly. The real question is how long they should suffer for their past crimes. I think 20 years is good, so I'd continue punishing them for another ten. Maybe require them to support a BSD lichened file system as an alternative to their own.

Speaking thereof, what woudl be a good open source alternative to exFat? One that has good properties for usb thumb drives and sd cards?

Reply Score: 2

Comment by marcp
by marcp on Tue 18th Sep 2012 23:22 UTC
marcp
Member since:
2007-11-23

I use ext2/3 and FFS/UFS on my SD cards and it works fine ;) but hey, I'm not a "multimedia" guy. I mostly run embedded stuff and such. I don't need it to be accessible by some fancy shmancy recorder/player, whose manufacturer was forced to pay some rediculous amount of money for some crappy exFAT support.

Reply Score: 1

RE: Comment by marcp
by Morgan on Wed 19th Sep 2012 03:53 UTC in reply to "Comment by marcp"
Morgan Member since:
2005-06-29

That's great for you, but (I mean no offense) you are the minority.

I think it would be awesome if camera manufacturers and other big players would all settle on ext2 or FFS for removable media, given the ease of licensing and better performance on the media itself. That would force Microsoft and Apple to incorporate support in their kernels as well, and we'd all be happier for it.

I'm not holding my breath though.

Reply Score: 4

RE[2]: Comment by marcp
by tidux on Wed 19th Sep 2012 14:26 UTC in reply to "RE: Comment by marcp"
tidux Member since:
2011-08-13

I'd say ext4, really. It's safer than ext2, and actually faster on flash media. It's only a tiny bit slower than FAT32 but can do silly things like hold a 4.7GB DVD iso.

Reply Score: 2

RE[3]: Comment by marcp
by Morgan on Wed 19th Sep 2012 14:33 UTC in reply to "RE[2]: Comment by marcp"
Morgan Member since:
2005-06-29

I thought journaling file systems were bad for flash devices though? That's why I was suggesting simple non-journaling file systems. I'm not a file system expert by any means though. ;)

Reply Score: 4

RE[4]: Comment by marcp
by Neolander on Wed 19th Sep 2012 15:29 UTC in reply to "RE[3]: Comment by marcp"
Neolander Member since:
2010-03-08

I thought journaling file systems were bad for flash devices though? That's why I was suggesting simple non-journaling file systems. I'm not a file system expert by any means though. ;)

I think the "journaling is bad for flash devices" part depends on one's priorities when it comes to mass storage. Personally, journaling has brought me so much benefits in terms of convenience and reliability in the past that I wouldn't want to live without it, no matter what the underlying hardware is.

After all, if modern flash drives couldn't handle journaling, they likely wouldn't survive constant writes to small user files either, which is exactly what many modern OSs do. The optimizations that are used for one, can likely be used for the other too.

Reply Score: 3

RE[4]: Comment by marcp
by libray on Wed 19th Sep 2012 17:25 UTC in reply to "RE[3]: Comment by marcp"
libray Member since:
2005-08-27

The journal is not the problem. The problem with the default mount options for ext2,3 and 4 are asynchronous writes in the name of speed. Mount all those disks with sync if you want safer.

This is all divergent from the article though. RIM, All Android makers, camera makers, and pretty much any device that reads an SD card must support FAT to interop.

Reply Score: 4

Comment by ilovebeer
by ilovebeer on Wed 19th Sep 2012 00:13 UTC
ilovebeer
Member since:
2011-08-08

The reason FAT is used on sd cards shouldn't be a surprise to anyone. I'm sure there's already an ancient whine-fest already posted here if you search.

And calling this `paying protection money`? Dial the drama queen'ing back a little, it sounds....dumb.

Reply Score: 5

RE: Comment by ilovebeer
by TechGeek on Wed 19th Sep 2012 02:56 UTC in reply to "Comment by ilovebeer"
TechGeek Member since:
2006-01-14

FAT as far as I know is patent free. Fat32 has a patent on it regarding the method of supporting the larger filenames. However, it can be programmed around (as is done in the linux kernel) and the ITC recently found prior art (Linus Torvalds) which should overturn the patent on it. The problem isn't the FAT file system. The problem is that Microsoft refuses to support any free file system which could be used. As such, they are leveraging their monopoly in the desktop market to affect another. That is grounds for anti trust actions. The easy and obvious way to avoid this would be for them to implement ext2/3/4. But then they wouldn't be able to sue over it.

Reply Score: 3

RE[2]: Comment by ilovebeer
by phoudoin on Wed 19th Sep 2012 08:15 UTC in reply to "RE: Comment by ilovebeer"
phoudoin Member since:
2006-06-09

The problem isn't the FAT file system. The problem is that Microsoft refuses to support any free file system which could be used. As such, they are leveraging their monopoly in the desktop market to affect another.


So true.
But their monopoly in the desktop market matters less than it used to.
Today, all major mobile OSes but Windows Phone are perfectly able to support free file systems out of box, making formatting your SD card in FAT less mandatory, except if you often move your SD card out of your phone to your desktop or you pro camera.

And this put Windows Phone in an unique position: being the late competitor, users wanting to switch to Windows Phone may ask more and more to be able to keep their SD card content from their current Android or iOS mobile device without having to reformat first them into exFAT, possibly forcing MS to actually support at least one free file system like ext2/3/4.

May I was Google, I'll push to format by default any new SD card inserted in an Android device to a free file system. Could be a good marketing operation too: see, we don't use patented/proprietary technology that could trap your private data.

That is grounds for anti trust actions. The easy and obvious way to avoid this would be for them to implement ext2/3/4. But then they wouldn't be able to sue over it.


Or the open source community to implement it for Windows plateforms. Someone will first have to cover the Windows FS kit cost, though.

An alternative could be to keep FAT32, but store only two files on it: a set of 4Gb "block" files hosting a guest free file system in it, and one single file, relying on auto-run feature: a free Windows tool to manage the guest file system within block files. Like the ZeroCD trick used by USB devices these days to embebded their drivers, but for file system.

Zero-FAT, or FAT-free, isn't that cool names for a technology!?

;-)

Edited 2012-09-19 08:23 UTC

Reply Score: 1

RE[3]: Comment by ilovebeer
by smashIt on Wed 19th Sep 2012 10:02 UTC in reply to "RE[2]: Comment by ilovebeer"
smashIt Member since:
2005-07-06

FAT at least up to FAT32 is patentfree
long filenames are patented, but you don't need to support them
in fact i haven't seen a single camera that uses something different than 8.3 naming

Reply Score: 4

RE[4]: Comment by ilovebeer
by phoudoin on Wed 19th Sep 2012 13:19 UTC in reply to "RE[3]: Comment by ilovebeer"
phoudoin Member since:
2006-06-09

FAT at least up to FAT32 is patentfree
long filenames are patented, but you don't need to support them
in fact i haven't seen a single camera that uses something different than 8.3 naming


Today SD cards aren't anymore used to store photos.
I'm seeing a pattern where the SD card is becoming to mobile devices what the USB keys are to desktop/laptop computers: removable storage.

In that context, it matters more to have better file system features set on the new removable storage support than when it was used only for music and photos.

And the only reason why [ex]FAT[32] is still this file system is for out-of-box interoperability with Windows users, and because the main use case was to make a copy of your data stored on your main Windows computer to be able to access (read, mostly) them from a mobile device.

This use case is moving toward a less desktop centric storage, via cloud and/or versatile compact storage support like SD cards and where the content can be generated directly from the mobile device.

The once mandatory file system seamless integration with Windows thru FAT can now be replaced with a software manager tool, as one could do with tarballs for instance.

My point is that even if Microsoft can forbid Windows to support natively all file systems but theirs, it's perfectly possible to use the plain old FAT with 8.3 names as a backend to a far more open, free guest file system that would launch automatically the software manager on a Windows system.

A FATPatent-free solution.

Reply Score: 2

RE[4]: Comment by ilovebeer
by Neolander on Wed 19th Sep 2012 14:48 UTC in reply to "RE[3]: Comment by ilovebeer"
Neolander Member since:
2010-03-08

FAT at least up to FAT32 is patentfree
long filenames are patented, but you don't need to support them
in fact i haven't seen a single camera that uses something different than 8.3 naming

IIRC, it's not even long filename support that is patented, but simultaneous support of long and short filenames.

Reply Score: 3

RE[4]: Comment by ilovebeer
by kovacm on Wed 19th Sep 2012 17:18 UTC in reply to "RE[3]: Comment by ilovebeer"
kovacm Member since:
2010-12-16

FAT32? Patents worth MILLIONS of dollars??

yeah, right...

It was overturned partly on the basis of 20-year old Usenet posts Linus Tordvals and an anonymous Atari hacker made when someone asked how to implement long filenames. The ideas they sketched up off the top of their heads was what FAT32 did. i.e. It was obvious to someone skilled in the field.

journal articles on subject: http://www.wired.com/wiredenterprise/2012/03/ms-patent/

and source from 1992.:

https://groups.google.com/forum/?fromgroups=#!topic/comp.os.minix/0r...
https://groups.google.com/forum/?fromgroups=#!topic/comp.sys.atari.s...

Reply Score: 4

RE[3]: Comment by ilovebeer
by dsmogor on Wed 19th Sep 2012 12:57 UTC in reply to "RE[2]: Comment by ilovebeer"
dsmogor Member since:
2005-09-01

That's way too late. You'd have to re-flash billions of gadgets that have FAT support in their guts.

Reply Score: 3

RE[4]: Comment by ilovebeer
by phoudoin on Wed 19th Sep 2012 13:25 UTC in reply to "RE[3]: Comment by ilovebeer"
phoudoin Member since:
2006-06-09

Why reflash?
Publish a software manager with a migration feature that will turn your current FAT content into a FAT-hosted file system.

With some tricks, you can possibly make the new hosted file system easy to detect on systems with native support, making it mountable directly.

Reply Score: 1

RE[3]: Comment by ilovebeer
by vaette on Wed 19th Sep 2012 14:10 UTC in reply to "RE[2]: Comment by ilovebeer"
vaette Member since:
2008-08-09

Or the open source community to implement it for Windows plateforms. Someone will first have to cover the Windows FS kit cost, though.

This happened over a decade ago actually, you can get a fully functional ext2 driver for Windows here: http://www.fs-driver.org/

It is really an excellent piece of work, all features one would expect work: UTF-8 filenames, plugging into Windows disk caching system, you handle the partitions in the normal Windows disk tools, you can have the pagefile on ext2, it supports the more advanced indexing of ext3, and it is in general lightning fast. I always point this driver out when people complain about interfacing with NTFS from Linux, since using ext2 in Windows is really a lot easier. The only issue is the lack of journaling, but to be honest it is not an all that big deal for a desktop PC.

Reply Score: 4

RE[4]: Comment by ilovebeer
by phoudoin on Wed 19th Sep 2012 20:11 UTC in reply to "RE[3]: Comment by ilovebeer"
phoudoin Member since:
2006-06-09

Great.
Let's reformat our sdcard into ext2/3 and install this driver on our Windows machine and bye bye exFAT patent trap.

Reply Score: 0

Simple Cure
by Lorin on Wed 19th Sep 2012 01:02 UTC
Lorin
Member since:
2010-04-06

Attack patents at the source, the patent office, get any refusals to the Supreme Court and most will go away.

Appeals courts have already ruled to various degrees that software is an abstract and as such can't be patented, the Supreme Court refused to hear challenges stating that they could not be overcome

Reply Score: 2

Nothing wrong with this
by vaette on Wed 19th Sep 2012 08:11 UTC
vaette
Member since:
2008-08-09

See, this I don't mind all that much. exFAT is a real specification designed entirely by Microsoft in recent times (first released 2006). I am not the least bit surprised that Microsoft has non-trivial intellectual property in it. On the other side, RIM is not paying up because they are using some trivial technique that happens to fall under a software patent, they are paying up because they are specifically implementing exFAT exactly as specified by Microsoft.

The only iffy part is that exFAT is so attached to a hardware standard in turn, but that does not seem to be the part people complain about.

I do think that the patent system needs to be reformed and reconsidered, but this is clearly not "paying Microsoft protection money", this is paying for the right to use a rather complete product that Microsoft designed.

Reply Score: 4

v Stamp Tax
by Snial on Wed 19th Sep 2012 11:55 UTC
Re:
by kurkosdr on Wed 19th Sep 2012 12:00 UTC
kurkosdr
Member since:
2011-04-11

Sorry, but this is open source paying for it's sins. If the open source folks had the smarts to replace the broken X.org in their OSes (be it Linux, PC-BSD or OpenIndiana) and made their OSes usable*, Windows wouldn't have 90+% marketshare, and Microsoft wouldn't have the power to essentially mandate exFAT for SDXC cards (exFAT and NTFS are the only filsystems supported by Windows that allow files larger than 4GB).
But with open source OSes being under 1%, if Microsoft wants to crush ext3 and other royalty-free filesystems, they have all the power to do it.

And if you think I am trolling, I really don't care...

*=and by "usable" I mean not breaking compatibility with existing apps every now and then and not breaking upgrades

Reply Score: 1

Protection Money Really?
by lucas_maximus on Wed 19th Sep 2012 13:01 UTC
lucas_maximus
Member since:
2009-08-18

Microsoft Corp. announced today that Microsoft and Research In Motion (RIM) have signed a patent licensing agreement that gives RIM broad access to the latest Extended File Allocation Table (exFAT) for certain BlackBerry devices of RIM.


That isn't hyperbole Thom, that is just a statement. The rest might be guff, but that is a pretty clear statement.

The conversation probably went like this down RIM:

* RIM Engineer - "We want to use exFAT, and that requires a volume license."
* RIM Manager - "Well lets get one, then."
* RIM Manager to Microsoft - "how much is an exFAT license?"
* Microsoft - "It is $X for this number of devices".
* RIM Manager - "That is with-in the budget, I will put you in contact with accounts."

FFS, This is how businesses normally work.

Edited 2012-09-19 13:02 UTC

Reply Score: 5

RE: Protection Money Really?
by dekernel on Wed 19th Sep 2012 14:14 UTC in reply to "Protection Money Really?"
dekernel Member since:
2005-07-07

If I had to guess, I would think you are spot-on.

Reply Score: 1

RE[2]: Protection Money Really?
by lucas_maximus on Wed 19th Sep 2012 16:34 UTC in reply to "RE: Protection Money Really?"
lucas_maximus Member since:
2009-08-18

I work in the Gambling Sector and it pretty much boils down like that. I think personally people forget or just don't know how a project works in business, a budget is assigned based on the requirements and cost estimates.

If exFAT support wasn't needed or out of budget RIM, they would have just supported FAT32.

Also people forget if a department doesn't spend it Budget it loses it.

Edited 2012-09-19 16:40 UTC

Reply Score: 3

RE[3]: Protection Money Really?
by moondevil on Thu 20th Sep 2012 18:02 UTC in reply to "RE[2]: Protection Money Really?"
moondevil Member since:
2005-07-08

Yeah, when I started working for a big multinational, I finally understood how big companies see open source and licenses of closed source software really work.

In the end it all boils down to available budget, possible license issues, the target demographic of the application and how to get the investment paid back.

In the corporate world there is no place for language or operating system religion, specially in the gaming sector, where the goal is to make games, regardless of what tools are available in the target platform.

Reply Score: 3

lucas_maximus Member since:
2009-08-18

I used to have a very idealistic ideas about Software etc, and tbh I changed while working because of exactly what you said.

You gotta get stuff done, and it has to be done to a budget and any other reasons management aren't interested in are just dismissed.

It was a harsh learning experience for me, made me think differently and I think it is easy for tech bloggers to say things and not really consider what is actually happening in a company.

Reply Score: 2

How silly Thom!
by jefro on Wed 19th Sep 2012 19:34 UTC
jefro
Member since:
2007-04-13

It is not protection. RIM for some reason needed to have extra code and full use of exfat for their product or series of products. So why not pay for what you need? That isn't against civil ways to live.

Do you pay protection money to your government when you pay taxes. Well, maybe but you also get services like fire and police and sewers and laws.

Same here. They claim to have paid for a broad range of technology, not just the right to use a filesystem.

Reply Score: 1

RE: How silly Thom!
by Alfman on Wed 19th Sep 2012 20:09 UTC in reply to "How silly Thom!"
Alfman Member since:
2011-01-28

jefro,

"It is not protection. RIM for some reason needed to have extra code and full use of exfat for their product or series of products. So why not pay for what you need?"


It's *possible* that RIM genuinely gained some MS technology in the deal, but I wouldn't say it's a forgone conclusion. Do you know for a fact that RIM needed or wanted MS code? Or are they paying royalties over functionality that their own engineers have already implemented themselves?

You might just call it a business transaction and leave it at that, but all too often these patent licensing deals are about avoiding lawsuits and have nothing to do with the technology.

Reply Score: 3

RE[2]: How silly Thom!
by lucas_maximus on Thu 20th Sep 2012 09:34 UTC in reply to "RE: How silly Thom!"
lucas_maximus Member since:
2009-08-18

http://www.osnews.com/permalink?535694

This is how business works my friend.

Reply Score: 2

RE[3]: How silly Thom!
by Alfman on Thu 20th Sep 2012 14:24 UTC in reply to "RE[2]: How silly Thom!"
Alfman Member since:
2011-01-28

lucas_maximus,

"This is how business works my friend."


I hope you understand the difference between acknowledging how business works, and justifying how business works. A corrupt official might say the same thing when accepting a bribe.

Reply Score: 2

RE[4]: How silly Thom!
by lucas_maximus on Thu 20th Sep 2012 16:47 UTC in reply to "RE[3]: How silly Thom!"
lucas_maximus Member since:
2009-08-18

Did you read my comment I linked?

Most businesses medium and larger businesses decide whether it is worth going into the deal.

Recently I recommended we buy a font-deck subscription, because we use Myriad, Bliss and a few other fonts that aren't freely available. Font deck was cheaper than buying the font outright and we needed to comply with the Copyright of the font ... Management agreed it was worth the extra money.

Microsoft have a patent (whether you agree with whether there should be a patent awarded for this is a different matter) and if you want to use exFAT in a product you are selling you should pay to use it.

A deal was done, and there was no court room dramatics or any other sillyness and there was simply an announcement that Microsoft made a deal with RIM.

Sorry whether you agree with software patents or not, nothing illegal nor immoral happened.

Edited 2012-09-20 16:53 UTC

Reply Score: 2

RE[5]: How silly Thom!
by Alfman on Thu 20th Sep 2012 20:20 UTC in reply to "RE[4]: How silly Thom!"
Alfman Member since:
2011-01-28

lucas_maximus,

"Did you read my comment I linked?"

Oh lucas, I was hoping the meaning of my response would be self evident, since it was also addressed to your original comment.


"Most businesses medium and larger businesses decide whether it is worth going into the deal."

It's true, deciding whether to pay for patents is a cost-benefit analysis. Ie: What are the royalties being asked for versus what is the risk and projected cost of litigation? But the mere fact that they get to make this choice does not automatically make it acceptable. Consider that companies paying mofia bribes also get to make a choice as well. Of course a principal difference is that they resort to physical threats, whereas patent holders resort to economic ones. But never the less both payments are a kind of insurance against the corresponding threats. This is surely what Thom was referring to by the term "protection money".



"Recently I recommended we buy a font-deck subscription... Management agreed it was worth the extra money."

There's nothing wrong with that, however you shouldn't be using a copyright example when we're talking about patents because they're not at all the same thing. Your company paid money for the right to use a 3rd party copyrighted font in it's own products, this saved your company the effort of creating it's own fonts. With patents, your paying for the right even to use your own private implementation.



"A deal was done, and there was no court room dramatics or any other sillyness and there was simply an announcement that Microsoft made a deal with RIM."

If the cost-benefit analysis pointed in microsoft's favor, this is what we'd expect.


"Sorry whether you agree with software patents or not, nothing illegal nor immoral happened."

I'm not disagreeing with you on the fact that software patents are currently our reality. But in reading your comments I get the impression that you think acknowledging our situation implies condoning it. Let me highlight why that's not the case:

Person A: acknowledge software patents and consider them morally acceptable.

Person B: acknowledge software patents, and consider them morally wrong.

Person C: won't acknowledge software patents at all.

The difference between (A-B) and C is a matter of facts. However the difference between A and B is a matter of opinion. In far fewer words, this is what was meant in my earlier post, and it applies to your original comment as well.

Reply Score: 2

RE[6]: How silly Thom!
by lucas_maximus on Fri 21st Sep 2012 07:24 UTC in reply to "RE[5]: How silly Thom!"
lucas_maximus Member since:
2009-08-18

fair enough

Reply Score: 2