Microsoft Expands exFAT Multimedia Licensing

alphadogg writes "Microsoft Thursday announced a broadening of its licensing program around its exFAT file system, which is designed to handle large multimedia files. Microsoft hopes companies making devices such as cameras and smartphones will adopt the Extended File Allocation Table (exFAT) technology to support the sharing of audio and video files. The technology is available on Windows 7, Vista SP1, Windows Server 2008 and Windows Embedded CE."

Comments

Bill's Ex is fat!

[Anonymous Coward]

Ballmer too...

I wish...

[mb1]

...I was exFAT.

EEE

[TheUni]

Embrace

---- You are here ----
Extend

Extinguish

(Thanks slashdot formatting-filter for making me sacrifice my ascii art skills.)

Re:EEE

[mk_is_here]

EEE only applies to open standard Microsoft targets. And exFAT is a proprietary standard solely developed by Microsoft. What is the point for ruining their own standard and make their customers and manufacturers angry?

Re:EEE

[rtfa-troll]

EEE only applies to open standard Microsoft targets.

It also applies to Microsoft partners. The multi-media product manufacturers (including cameras, media players etc. etc.) will be the long term target. Right now their functionality is being extended with the aim of Microsoft getting lock in. Microsoft is already one of them (with it's Windows Mobile phones and XBox at least). Later, when they need to expand their market, they will wipe out the multi-media companies that have become locked in.

The thing is, and I know this from working in a potential victim company and discussing with the person who was negotiating with MS for media standards, that the extinguish is at least five years away. Almost nobody working in such a company cares about that far in the future.

Only companies, like Oracle, which decide to fight Microsoft from the beginning as hard as they can, will ever survive long term in such a market.

Re:EEE

[jimicus]

It also applies to Microsoft partners. The multi-media product manufacturers (including cameras, media players etc. etc.) will be the long term target. Right now their functionality is being extended with the aim of Microsoft getting lock in. Microsoft is already one of them (with it's Windows Mobile phones

That's true, but outside the operating systems market Microsoft have found it very hard to achieve anything like the same amount of influence. Windows Mobile, for instance, just seems to get weaker and lose market share every year.

Re:EEE

[Tubal-Cain]

Windows users won't notice. Mac and *nix users, on the other hand...

I smell DRM

[Barryke]

I smell DRM.

Re:I smell DRM

[bogaboga]

You are just part of a bunch of folks who are apprehensive about whatever Microsoft does...even in good faith.

I do not want to come across as judgmental but I am sorry you just seem to be. Why do you always look at Microsoft will "all" the disdain? Why?

Microsoft is in the business of making money and licensing of its wares is just part of the game. What's wrong with that? Did you want Microsoft to go the Linux way and "donate" the software for "free?"

Get a life...Have some faith.

Re:I smell DRM

[Anonymous Coward]

I smell embrace, extend, extinguish.

Re:I smell DRM

[Opportunist]

Why do you always look at Microsoft will "all" the disdain?

Experience.

Fool me once, shame on you. Fool me twice ... umm... you won't get fooled again.

Snide comments aside, it's simply that it's been too many times the case. Of course MS is in the business to make money. But to that end, vendor lock-in is one of the golden tickets to cash cows. If you can monopolize, you can charge whatever you want and nobody can undercut you. You can dictate price, conditions and format, what your user may or may not do with your tools and so on.

Yes, MS is in the business to make money. And doing what we "accuse" them to do is the easiest, most profitable and most sustainable way. So I guess we might be correct?

Re:I smell DRM

[Anonymous Coward]

exFAT has little to do specifically with audio/video - it's a filesystem designed for flash memory. TFA is misleading in this regard.

Re:I smell DRM

[Anonymous Coward]

No, it's because everytime I come near a machine running windows, I lose around half of my nerves from all the

-EULAs,
-"Yes, please send all my personal data to Microsoft and/or McAffee, Delle, Evilcporp"-checkboxes,
-"YOU ARE A THIEF UNLESS YOU PROVE OTHERWISE"-'advantage'-dialogs,
-"Logging off"-Screens that stay on for 10 minutes before the computer shuts down
-"There are important updates", "there are more updates, "are you sure you want to do this", "I have a message" "ballons"
-"No you can't connect more than a single person via RDP, it's not allowed"
-"what do you mean 'virtualize'? Do you have a license for that ?"
-"Updating media player? Give us access to all your stuff"
-"non-localized EULAs"
-crashes
- Start your computer? Wait 15 minutes to go through through some updates.
- "Your Visual Studio 'forgot' how to display the search/replace dialog ? Why don't you reset all the personalizations you made, or start with a fresh profile?"

At this point I decided that I that really valued my sanity, and quit my job developing for .NET.

"Have some faith he" said...
"what's wrong with licensing?" he said..

"If we wanted Microsoft to 'donate' their software for 'free'" he asked.
No. Not really, me personally I am trying not using their stuff, and I am successfull not even being anywhere near it atm.

However I want Microsoft to 'donate' for 'free' dataformats they are using so that all of us could exchange data, and if we didn't like an application or one wasn't available for our platform that was using said format, we could write our own. And maybe if it was better than theirs they could say, "Hey look at what that bloke is doing, maybe if we would ask nicely, he would let us integrate parts of his functionality into our own application", and then I would say "Yes sure, I was also liking that one feature there, could I maybe also...? " and then they would say, "yes sure, why don't we invite this developer over there as well ?"
And then we could develop the stuff we need together and get to do much more of the dataprocessing done that we actually want to do, instead of fighting over stupid licenses or writing complicated copyprotection-software that never works.

A man can dream....

Re:I smell DRM

[Stormwatch]

Are you serious? I wonder, have you ever heard of:

* The AARD code?
* OOXML?
* The Halloween documents?
* Embrace, extend and extinguish?
* Samizdat?

"Have some faith", you say? Indeed, to trust Microsoft to act ethically is a matter of faith: to believe in something incredible against all evidence.

Re:I smell DRM

[PenisLands]

Microsoft does not do anything in 'good faith'. In this particular case it's obvious what Ms is trying to do - implement a proprietary filesystem which only their (most recent) OSes support, then get all the device makers to use it. Linux, Mac, and XP are left out in the cold and users must pay up for Windows 7 in order to use their new camera/whatever.

Then if it suits them, they might try to use copyright/intellectual property/patent claims to keep Linux from implementing it. Or threaten Linux using companies who are using the Linux implementation.

Re:I smell DRM

[GigaplexNZ]

You had me until "crashes". All software is vulnerable to bugs, and bugs aren't intentional.

Re:I smell DRM

[Nerdfest]

That's right, we 'Won't get Fooled Again' ... hmmm

*Takes sunglasses off, a la Caruso*

It's time to ... cut the exFAT

Yeeeeeaaaaahhh!!!!!!!!!!

Re:I smell DRM

[pesc]

...Microsoft is in the business of making money...

No, that is a secondary goal. The first priority for Microsoft is control, technology ownership and monopolization. Even at a financial loss.

See IE, XBOX, dotnet, Silverlight, etc, etc

Re:I smell DRM

[Anonymous Coward]

I smell stupid. Yes, its you.

Re:I smell DRM

[140Mandak262Jamuna]

Microsoft is in the business of making money and licensing of its wares is just part of the game. What's wrong with that? Did you want Microsoft to go the Linux way and "donate" the software for "free?"

Get a life...Have some faith.

Well, what is wrong in the customers resisting the profit motivated actions of their vendors? Customers have as much right to protect their money as does Microsoft have for making their profits.

Some actions of the vendors, including Microsoft, enhances the productivity and competitiveness of their customers. Rightfully the vendors, including Microsoft, are entitled to a share of the extra profits generated. But some other actions by the vendor, does not enhance the productivity or competitiveness of their clients, and the customer would be better served by switching to a competitor of the current vendor. Actions by the current vendor that prevents this switch by vendor lock would hamper the clients from employing their money, maximizing their profits etc. And we have as much right to highlight to potential long term danger and make everyone aware of it.

Why is Microsoft and its apologists are so against people making informed decisions? Vendor lock is real. Companies are hurting from it.

Re:I smell DRM

[krelian]

I must say that you really
*puts on sunglasses*
butchered this one.

Re:I smell DRM

[Anonymous Coward]

you can charge whatever you want and nobody can undercut you. You can dictate price, conditions and format, what your user may or may not do with your tools and so on.

Hmm, now what company does this sound like...

Re:I smell DRM

[Zontar The Mindless]

You left out DOS Ain't Done Till Lotus Won't Run.

Re:I smell DRM

[Stormwatch]

Except, for once, MS is innocent. They did many nasty things, no question, but that Lotus thing is probably a myth.

Re:I smell DRM

[jedidiah]

They stand to be a victim of this nonsense too.

In order to interact with the "monopoly product" everyone gets on board with a
vendor-lock storage solution that gives Microsoft the excuse to sue anyone else
that "dares to be compatible". The Monopolist represents the single largest group
of customers, so all the other 3rd parties gladly go along since it appears to be
in their immediate best interests.

This BS has been going on since it was MS-DOS vs. Macintosh.

Re:I smell DRM

[BigDXLT]

Yeaaaaaahhhhhhhh!!!

Re:I smell DRM

[Anonymous Coward]

No, that is a secondary goal. The first priority for Microsoft is control, technology ownership and monopolization. Even at a financial loss.

This is just stpid. Of course the first priority is money! It's just that, sometimes, a short-term loss (even a large one) is justified to further control - but not for the sake of control itself, but solely for greater money-making opportunities that it entails.

Re:I smell DRM

[sjames]

When someone (or something) does enough wrong, it comes to be expected. Even to the point where someone who doesn't expect the worst will be considered a fool.

If Jim Jones offers me koolaid, I'm going to decline. If Bernie Madoff offers me a 'sure thing' investment opportunity (from jail), I'll say no.

When a corporation that has been repeatedly convicted in multiple countries for shady business practices offers me ANYTHING, I'm going to at least have to look sideways at them. Especially when they have a long history of stabbing their business partners in the back. There's a good chance they're not merely looking for a mutually beneficial business transaction.

Why WOULDN'T a person at least be apprehensive of dealing with such an entity?

Re:I smell DRM

[Anonymous Coward]

"You are just part of a bunch of folks who are apprehensive about whatever Microsoft does...even in good faith."
          It's NOT in good faith. They are leveraging their monopoly status in OSes (they ARE twice-convicted, by US and EU, at least...) to get money from other markets for not actually creating anything new. (ExFAT involves no new technologies, just enough changes so they can consider it patentable.)

"I do not want to come across as judgmental but I am sorry you just seem to be. Why do you always look at Microsoft will "all" the disdain? Why?"
          Because everything, EVERYTHING, they've done in the past has been to screw over competitors or even "partners". They do not play nice. The more you look, the more bad deeds you'll dig up, going back to the 1980s.

"Microsoft is in the business of making money and licensing of its wares is just part of the game. What's wrong with that? "
          See the above -- they are trying to force people to pay them money, when they have come up with nothing new. And just building it in preemptively into Vista, 7, etc. as a leverage to prevent the camera makers from just coming up with an *appropriate* filesystem to use.

Why?

[Teun]

Why, when you can pick up ext2 for free?

Re:Why?

[bucky0]

To be able to read it on almost every computer available? There's a benefit to that when you have removable media, you know.

Re:Why?

[Teun]

You'll need to develop and install a driver anyway, why not take an existing one that's unencumbered?

Re:Why?

[hitmark]

simple, the only non-microsoft formats that windows supports out of the box are cd and dvd media.

i wonder how long it will take before microsoft gets a slap on the wrist over this...

new microsoft, same as old microsoft...

Re:Why?

[peppepz]

What about UDF? It's already supported out of the box on removable media by Windows Vista and higher.
Open standard, tons of features, fast on flash media, broad adoption by existing operating systems and devices.

They should use it instead of inventing yet another file system with less features. And closed, too (so much for Microsoft's commitment to interoperability and open standards).

Re:Why?

[oglueck]

Well, the CD-ROM standard they support is "Joliet". Which is their own extension.... I wonder how long until they are going with patents after others implementing it.

Re:Why?

[Renegrade]

Almost every computer available? Hardly. From the article: "The technology is available on Windows 7, Vista SP1, Windows Server 2008 and Windows Embedded CE". That's it. The Win 98, 2000, and XP systems you'll find in the wild won't support it. Some of the older systems (ie, XP) can be patched with an update from Microsoft, but are you going to carry a second removable media device with FAT16 or FAT32 around with you and install this patch everywhere you go? And bring XP or later as well for those machines running 98/NT4/2K? I don't believe there's Apple support either, and Linux support is still experimental.

I haven't seen the spec for exFAT (I'm not paying some fee to see a spec for some microsoft cruft), but I imagine it's another vendor-lockin, poor-performance-substitute abomination like NTFS was, or WinFS will be.

Re:Why?

[Anonymous Coward]

Says who? Most media from cameras work without drivers.

Re:Why?

[Anonymous Coward]

He was talking about a driver for the file system, not the device. exFat requires one becuause it's not supported out of the box on anything but Microsoft's newest OSes.

Re:Why?

[bertoelcon]

UDF is also used on DVD Media. That is why it is supported.

Re:Why?

[peppepz]

Of course! But starting from Vista, it's supported on all other media, too. So what's the need for exFat? I think UDF might cover all of exFat's use cases, with no patents pending and secret specifications involved.

Re:Why?

[Anonymous Coward]

Some of the older systems (ie, XP) can be patched with an update from Microsoft, but are you going to carry a second removable media device with FAT16 or FAT32 around with you and install this patch everywhere you go?

Never mind the computers you can't patch. At the moment exFAT would be completely useless to me, because my own computers don't support it (I run Linux) and because all the other computers I use -- at university and its library -- don't support it either (still running XP; they might get the patch at some point though, but most likely it's going to take a while).

Re:Why?

[Jugalator]

WinFS? Really? :p

Re:Why?

[A Friendly Troll]

Well, the CD-ROM standard they support is "Joliet". Which is their own extension.... I wonder how long until they are going with patents after others implementing it.

ISO-9660 doesn't support Unicode. Believe it or not, some languages use characters that aren't part of ASCII.

ISO-9660 doesn't support lower case letters, spaces and multiple dots in file/directory names.

There's nothing wrong with naming a directory "Family Photos 25.12.2009." - if Joliet didn't exist, we'd have to burn that to CD as "FAMILYPHOTOS25122009".

Re:Why?

[TheRaven64]

Starting with Vista is the problem. There are still lots of XP machines around that would like to be able to write to memory cards from various devices (I think they can read from UDF already). OS X only got write support for UDF quite recently, so there are probably a lot of Mac users without it too. At the very least, you'd need to provide an IFS for XP that properly supported UDF. You could port a lot of it from FreeBSD, but it still might be cheaper just to license the MS patents.

Re:Why?

[Anonymous Coward]

if Joliet didn't exist, we'd have to burn that to CD as "FAMILYPHOTOS25122009".

You picked a bad example: you can create such a directory using RockRidge.

Your point is sort of correct however. Joliet is to Windows as RockRidge is to UNIX. Microsoft needed some way to support DOS style attributes and stuff like FAT32 short/long filenames, and neither ISO9660 nor RockRidge offered that support, so they created Joliet. Nothing much wrong with that.

Re:Why?

[hitmark]

and to add insult to unjury, i think microsoft have already provided xp with a exfat patch that provides full read and write support...

Re:Why?

[Anonymous Coward]

There's nothing wrong with naming a directory "Family Photos 25.12.2009." - if Joliet didn't exist, we'd have to burn that to CD as "FAMILYPHOTOS25122009".

Wrong. You COULD use rock ridge -- http://en.wikipedia.org/wiki/Rock_Ridge . Of course microsoft prefers to come up with its own thing.

Re:Why?

[DrXym]

The problem there is that ext2 is as associated with Linux as FAT32 is with Windows. But it doesn't enjoy the ubiquity of FAT32 and probably never will. Even though ext2 drivers do exist for Windows, who is going to bother installing them to read large content from a flash drive?

The best chance ext2 / ext3 / ext4 or any other fs has for multimedia storage is when the user doesn't even know or care what file system is being used. I wouldn't give a damn what fs a NAS / streaming devices is using so long as it works and is reliable. I'm sure that exFAT would be largely redundant if someone defined a simple, clearcut way to break up large files into 4Gb chunks for easy transfer over FAT32.

Re:Why?

[Malc]

NTFS is actually a very good FS. The performance isn't so bad for desktop usage at least, and I've hard far fewer problems with it over the years than with say ext2/ext3 from a corruption point of view. It's easier and more reliable to read on my Mac (or on Linux boxes) than ext2/ext3 is (I had to uninstall the OS X drivers due to issue they caused, and the Windows support is a joke). The only downside is that nobody seems to be able to create a driver with reliable write support.

Re:Why?

[magamiako1]

NTFS has too much overhead for removable media devices, which is why they created exFAT (FAT64).

Re:Why?

[A Friendly Troll]

Wrong. You COULD use rock ridge -- http://en.wikipedia.org/wiki/Rock_Ridge . Of course microsoft prefers to come up with its own thing.

Yes, or I could use the Apple extensions, or the Amiga version of Rock Ridge, or god knows what else...

We are talking about Windows here - perhaps I haven't been clear enough.

Non-Windows systems have never required Joliet, but Windows has, because all other solutions are technically inadequate for that OS. For that very reason, Joliet is technically inadequate on MacOS and UNIX-like systems, the Apple ISO-9660 extensions are inadequate on UNIX, etc, etc.

Now, if you want to propose a unified spec that would be technically adequate for all operating systems and all file systems, then by all means, go ahead. In the meantime, Joliet is a technical necessity for Microsoft OSs.

Re:Why?

[vrmlguy]

Non-Windows systems have never required Joliet, but Windows has, because all other solutions are technically inadequate for that OS.

In what way is Rock Ridge "technically inadequate"? For that matter, why not just use UDF? It's designed for all optical media, not just DVDs, and has supported Unicode for almost a decade.

Re:Why?

[0ld_d0g]

but I imagine it's another vendor-lockin, poor-performance-substitute abomination like NTFS was, or WinFS will be.

You can believe that if you want. You have to show evidence for it though if you want others to believe it. What don't you like about the algorithms used in NTFS? Feature wise I cant think of any major issues in NTFS. http://en.wikipedia.org/wiki/Comparison_of_file_systems#Features

Traditionally, Windows(NTFS) has always been good at random access I/O and Linux(ext2/3) at sequential I/O. My experience has been ext2/3 always lagged behind NTFS in handling large files (>3GB)

Re:Why?

[A Friendly Troll]

In what way is Rock Ridge "technically inadequate"?

Take a guess? Also, does it support Unicode?

For that matter, why not just use UDF? It's designed for all optical media, not just DVDs, and has supported Unicode for almost a decade.

UDF (consumer-level) is plagued with incompatibilities, not just on computers.

Re:Why?

[MrMr]

I think UDF might cover all of exFat's use cases, with no patents pending and secret specifications involved.
So where's the profit in that?

Re:Why?

[Anonymous Coward]

Rock Ridge is for POSIX-style filesystems. It allows for all kinds of things that Windows wouldn't be able to handle. For example, any character apart from a null or / character is allowed in a filename. Windows doesn't allow any of \ / : * ? " or |

Similarly, Rock Ridge allows for hard links (which only work on Windows NT, and not the same way they do on POSIX systems), symbolic links (not supported until Vista, and nothing like the POSIX ones), device nodes (don't exist outside of the Windows NT kernel, and can't exist on a real filesystem), unix-style ownership and permissions (which make no sense in any version of Windows), POSIX semantics for timestamps (UTC instead of local time, and stored as a Unix timestamp), and much deeper nesting of directories beyond the 260-character path limit in Windows. Oh, and filenames are case-sensitive.

Admittedly, most of those limitations apply more to Windows 9x than Windows NT, but many of them still persist in all applications that use the Windows API, particularly the filename and path restrictions. Still, a Rock Ridge filesystem doesn't really make any sense on Windows, it's nearly impossible to support fully on current versions of Windows, and was completely impossible to support fully on Windows 9x.

Conversely, Joliet is basically ISO9660, with a few of the restrictions relaxed to match those of Windows (path length, directory nesting depth, and allowed characters), and filenames encoded as UCS-2. Critically, it supports Unicode character sets (using an obsolete encoding, of course), which Rock Ridge does not (you can use UTF-8, but it's not really part of the standard - it just happens to work).

Joliet works OK on non-Windows systems, and it's possible to fully support (assuming you convert the filenames to UTF-8), but doesn't act anything like a POSIX filesystem. On a Mac (particularly pre-OS X versions), it would have been impossible to store applications or files that used the resource fork, so straight ISO9660 or Joliet are not very useful, hence the Apple-specific RR extensions, and the hybrid ISO9660/HFS format.

UDF would have been a decent choice for modern systems, but it seems to have never really caught on. Personally, I'd blame Windows XP for this - it never supported UDF on anything other than optical drives, and even then, packet writing required separate software which provided it's own UDF driver anyway.

Re:Why?

[Waffle Iron]

What don't you like about the algorithms used in NTFS?

I don't like that they're shrouded in a black box.

Re:Why?

[Petaris]

I've been thinking the same thing for a long time. FAT really does need to be replaced on removable media with something that is better (larger partition sizes, handles large files better, doesn't fragment). EXT2 or 3 would be good options. I can't believe that MacOS X doesn't have native EXT2/3 support though. And the drivers people have created for it are not the best, though I applaud the attempts.

Re:Why?

[Megane]

I can't believe that MacOS X doesn't have native EXT2/3 support

Blame the GPL license. Most of OS X is under a BSD-style license. Is there even a BSD-licensed EXT2/3 implementation?

There is a pre-built version that you can download and install yourself, however.

Also, the only people who need it are the 2% who use Linux. 2% of 5% isn't much, especially when OS X users have much less need to keep Linux around than Windows users.

And we OS X users already have a very nice filesystem, thank you very much. Apple did add an amazing number of hacks to it so that it can do Unix-y things like inodes and hard links, but it works very well. It can even be made case sensitive should we ever want to compile Linux. (A million curses on whoever required case sensitive filenames in the netfilter code.)

Re:Why?

[donaldm]

The UDF file-system is for Optical Storage not magnetic or even solid state storage. So this is not really a contender.

Re:Why?

[mishehu]

Why ext2? How about UDF? I know it was originally intended for optical media, but it works quite well for non-optical media. I use it on a number of flash drives that I have that I want to use between Windows and Linux systems...

Re:Why?

[10101001 10101001]

ISO-9660 doesn't support Unicode. Believe it or not, some languages use characters that aren't part of ASCII.

That probably has something to do with ISO-9660 being started as High Sierra in 1985 and the first published standard appearing in 1988 while Unicode was first started in 1987 and the first published standard appearing in 1991.

ISO-9660 doesn't support lower case letters, spaces and multiple dots in file/directory names.

You know, those limitations sound remarkable like MS-DOS limitations. Oh, right, to provide compatibility between different OSs (Unix, Windows, and Macs), ISO-9660 was designed around a few lowest common denominators, like short limited character filenames.

There's nothing wrong with naming a directory "Family Photos 25.12.2009." - if Joliet didn't exist, we'd have to burn that to CD as "FAMILYPHOTOS25122009".

If Microsoft had chosen a more sensible scheme for FAT filenames in the first place, ISO-9660 would almost certainly have had support for spaces, multiple periods, and perhaps even lowercase letters. This leads to the GP's point: Microsoft striving to patent solutions to problems it created and then actually collecting royalties on those solutions. But, yes, yippie that we have Joliet.

Re:Why?

[Anonymous Coward]

new NSA, old NSA

Re:Why?

[arkane1234]

My experience has been ext2/3 always lagged behind NTFS in handling large files (>3GB)

My experience has been that you never have a single file that exceeds 2tb, which is the filesize limit on ext3.
3GB is a drop in the bucket... and this is from experience.

Re:Why?

[peppepz]

UDF can be used on every media.
For my flash media, under Windows, it performs better than FAT and exFAT.

Re:Why?

[shutdown -p now]

UDF doesn't magically know what kind of storage it is used on, and there's nothing special in its design that makes it unsuitable for use on magnetic or solid state storage. In fact, it fares somewhat better on SSD than many other general-purpose filesystems, because it's originally designed for media that wears out quickly.

Re:Why?

[Anonymous Coward]

I haven't seen the spec for exFAT (I'm not paying some fee to see a spec for some microsoft cruft), but I imagine it's another vendor-lockin, poor-performance-substitute abomination like NTFS was, or WinFS will be.

NTFS has everything in common with HPFS (also known as OS/2's file system) and almost nothing in common with FAT except the fact that they both work on magnetic disk drives.

NTFS's main problem is not performance per se, but the fact that most of the documentation for the format is unavailable unless you sign a waiver. Hence the issue of interoperability with Linux until NTFS-3G was released.

Re:Why?

[yuhong]

"are you going to carry a second removable media device with FAT16 or FAT32 around with you and install this patch everywhere you go?"
You can partition your flash drive.

Re:Why?

[afidel]

I think the fact that journaling and flash media don't mix particularly well is the main reason not to use NTFS.

Re:Why?

[yuhong]

Yep, NT once was supposed to be a version of OS/2, and the early versions of WIndows NT even supported HPFS filesystems. (Removed in NT 4.0)

Re:Why?

[Jesus_666]

I think UDF might cover all of exFat's use cases, with no patents pending and secret specifications involved.

Which is specifically why it won't be used.

Re:Why?

[badkarmadayaccount]

What if the flash drive shows up as a NAS?

Re:Why?

[Renegrade]

Ah, and which partition format will we be using that works with everything? And will this patch add these features to non-Windows boxes, or ones with restricted Admin rights, like most corporate machines?

Re:Why?

[Renegrade]

I've found that fat32 drives are faster and more reliable. The only downside being the 4G limitation and that Microsoft disapproves of the patent-resistant/DRM-proof design and lack of unnecessary complication, and thus artificially caps volume size at 32 gigs (not an issue under other OSes, including this old Win98 boot disk I created long ago).

I've never had any issue with ext2/3 personally, aside from the fact that it's a bit slow on massive file deletion.

Oh, fat32's lack of permissions, in a Windows environment, is actually a good thing. I don't know how many times Windows has managed to completely screw up share permissions on my NTFS shares...

Microsoft and Making Money

[ztransform]

Microsoft is in the business of making money and licensing of its wares is just part of the game.

Creating a software product and selling it is fair enough. Creating a standard, expecting everyone to use it, then charging a license fee for it is _evil_.

That's like the power company deciding to sell you power. Then charging you license fees for installing power sockets in your home that conform to the standard.

Re:Microsoft and Making Money

[TrancePhreak]

You are obviously not a certified electrician.

Re:Microsoft and Making Money

[Gadget_Guy]

Creating a standard, expecting everyone to use it, then charging a license fee for it is _evil_.

What exactly is evil? Firstly, they haven't created a standard. If they had then surely they would have published the specifications somewhere. exFAT is a proprietry file format.

I don't know if they expect everyone to use it, although they may hope that everyone uses it. If it is a sin to hope that your product is popular, then most of the companies in the world are going to hell.

Finally, why is charging a license fee for something evil? If you don't want to pay to use it, then don't use it. That is the same argument as saying if you don't want people to see your source code, don't incorporate GPL code into it.

Re:Microsoft and Making Money

[Tapewolf]

What exactly is evil? Firstly, they haven't created a standard. If they had then surely they would have published the specifications somewhere. exFAT is a proprietry file format.

I don't know if they expect everyone to use it, although they may hope that everyone uses it.

Everything that wants to SDXC will have to use exFAT. It's part of that standard. This is going to be inconvenient for anyone who wants to use their shiny new camera/camcorder on a Mac or linux netbook or someone else's XP machine.

Re:Microsoft and Making Money

[Tapewolf]

Ugh. Make that "Everything that wants to use SDXC."

Re:Microsoft and Making Money

[peppepz]

Since exFAT apparently is referenced in the SD standard, people will be forced to use it, if they buy any consumer electronic device containing an SD slot. They can't choose not to use it. It's a hardware standard.

So after exFAT, they won't be able to do what they do today, that is, freely exchange their media among their devices at their will. That's evil, and once again, it comes from Microsoft.

Re:Microsoft and Making Money

[Anonymous Coward]

Right, so the power companies own the electrician regulatory authorities?

Re:Microsoft and Making Money

[huge]

Everything that wants to SDXC will have to use exFAT. It's part of that standard.

Correct me if I'm wrong, but isn't SDXC defined by SD Association, not by Microsoft. Microsoft is one member of the association, I give you that, but there are several others as well. Unless Microsoft somehow coerced the association to select exFAT, I consider this to be a bad move by the association rather than Microsoft.

Re:Microsoft and Making Money

[Tapewolf]

Unless Microsoft somehow coerced the association to select exFAT, I consider this to be a bad move by the association rather than Microsoft.

That may be. It has the same net result though.

Re:Microsoft and Making Money

[abigsmurf]

They can choose to buy a camera which doesn't use SD cards. There are plenty of competing formats around. Sony would be more than happy for you to buy a Memory Stick using camera.

Re:Microsoft and Making Money

[Tapewolf]

They can choose to buy a camera which doesn't use SD cards. There are plenty of competing formats around. Sony would be more than happy for you to buy a Memory Stick using camera.

Unfortunately, Memory Stick XC also uses exFAT as its filesystem. See the exFAT (or Memory Stick) pages. Those (SD and Memory Stick) seem to be the main formats at the moment.

Re:Microsoft and Making Money

[Tapewolf]

Damn. Make that "See the exFAT (or memory stick) pages on wikipedia."

Re:Microsoft and Making Money

[vrmlguy]

Since exFAT apparently is referenced in the SD standard, people will be forced to use it, if they buy any consumer electronic device containing an SD slot. They can't choose not to use it. It's a hardware standard.

So after exFAT, they won't be able to do what they do today, that is, freely exchange their media among their devices at their will. That's evil, and once again, it comes from Microsoft.

Is there any reason why you can't use UDF on flash media? It's designed for media that wears out with too many writes, so it seems like a perfect fit. And recent (since ~2000) versions support Unicode, so you can use Tengwar Sindarin for your file names.

Re:Microsoft and Making Money

[melstav]

Yes, exFAT is part of the new SD specifications. So is FAT32.

What's really interesting about the exFAT specification is the terms of the license it's under (at least through the SD Association)

If you need access to the standard so you can build a device (or program) that will only need to understand the contents of the filesystem when it's accessed in a read-only manner, you don't actually have to pay a licensing fee. So, an MP3/4 player won't need a paid license for exFAT. It's only if the device needs to *WRITE* to the filesystem (such as a camera) that a paid license is required.

Re:Microsoft and Making Money

[Anonymous Coward]

> Unless Microsoft somehow coerced the association to select exFAT

Of course they did.

Microsoft do not support third-party file system drivers. It's certainly possible to write them, but documentation is scarce, there are few tools available, and it's intended for use only in embedded systems. It's not possible to provide the same level of integration as the Microsoft-provided drivers, since none of the disk management tools are extensible by anyone but Microsoft. The IFS drivers themselves are far more complicated than filesystem drivers in any other operating system, since they have to implement functionality that is normally provided by the operating system's VFS layer, and that functionality isn't well documented. Basically, all existing third-party IFS drivers suck.

The SDXC committee needed the memory cards to be usable on Windows, ideally without installing any drivers, and without having to screw around to get it to work. NTFS is completely unsuitable for flash storage, and is far too complex to implement in an embedded device anyway (not that Microsoft actually license any part of NTFS out to third parties anyway).

That leaves UDF, which can not be used on anything but optical discs in Windows XP, FAT16, which can't be used for drives larger than 2GB, and FAT32. FAT32 can actually scale well into the terabyte range, but Windows will refuse to format a disk as FAT32 if it's larger than 32GB.

Microsoft's solution was exFAT. The selection of exFAT instead of FAT32 was forced by limitations on Windows' FAT32 support. These limitations were intentional - the idea was to get people to use NTFS instead of FAT32. The only party to benefit from this situation is Microsoft - they get to sell licenses for their (apparently) patented new filesystem, which they can't do with FAT32 (cameras never used the patented VFAT extensions). In addition, they get to kill off one of their own filesystems, which was being used as a common interchange filesystem between completely different systems, none of which needed to run Windows, or use any Microsoft technology.

Re:Microsoft and Making Money

[nanoflower]

I don't see a problem with XP users and exFAT since Microsoft provides it for XP. http://support.microsoft.com/kb/955704 Now Mac and other OS users may have problems since exFAT isn't currently available for their OS but that will likely change should this file system take off.

Re:Microsoft and Making Money

[thejynxed]

Uh, if you don't plan on writing to the filesystem, how exactly do you propose anyone load music onto said media player if it only supports exFAT-enabled memory cards and the internal storage also relies on exFAT?

The manufacturers will end up paying licensing fees to Microsoft either way, for convenience if nothing else.

Re:Microsoft and Making Money

[Bert64]

Speaking of the intentional FAT32 limitations...
Try formatting a drive bigger than 32GB with fat32 on win2k, win98, linux or macos, it works fine..
With XP they crippled that functionality for no other reason than to force people to use the more proprietary ntfs.

Re:Microsoft and Making Money

[melstav]

You don't have to understand the filesystem to be able to write data to it, as long as something else is telling you where to put the bytes.

So, for example, an SD-Card reader doesn't have to have a paid license because the card reader doesn't have to understand the filesystem. All it has to know about is reading and writing blocks. The OS has to know how exFAT works if that's the filesystem on the media.

Re:Microsoft and Making Money

[melstav]

I should have appended to that comment an explanation of how the distinction applies to music players.

Say you have an MP3 player that uses exFAT. The player can read the files off the flash without needing a paid license.

Now, say you plug the player into your computer's USB port and it shows up as a USB-attached disk. (and/or card reader) The OS will need exFAT drivers to be able to copy files to/from it. But when the commands to access the filesystem reach the player, they're not "open suchandsuch file" they're "read/write block 123456".

Now, on the other hand, if the MP3 player included an FM tuner and the manufacturer wanted to allow you to record audio from the radio to the exFAT card, it would require a paid license because it would need to understand how to correctly write to the card.

Re:Microsoft and Making Money

[yuhong]

Actually it was 2000 that crippled this, and part of it is because of the time to calculate free space, according to a TechNet Magazine article by Raymond Chen.

Re:Microsoft and Making Money

[yuhong]

In fact, there is already an open-source read-only driver for exFAT on Linux.

Re:Microsoft and Making Money

[vegiVamp]

Waitwhut ? Are you telling me that it's impossible for me to format an SD card in something else than exFAT ? I can't remember ever trying, but I sure as hell am gonna try as soon as I get home.

Re:Microsoft and Making Money

[Tubal-Cain]

It's kind of pathetic, but I didn't notice the omission.

Re:Microsoft and Making Money

[pydev]

It won't be inconvenient for Mac users; Apple is Microsoft's lap-dog, their token competition. That's why Microsoft will license this to Apple one way or another.

Microsoft wants a duopoly, with them getting most of the market, Apple keeping the feds off their back, and no open source anywhere (or if there is open source, they want to control it through patents).

Re:Microsoft and Making Money

[Dishmopo]

No, the power companies don't own the electrician regulatory authorities; it's not necessary for them to. However, it would be pretty naive to think thay they had no influence over licensing requirements in the past.

And I can bet you they licensed their patents for electrical outlets as well.

Re:Microsoft and Making Money

[pydev]

Since exFAT apparently is referenced in the SD standard, people will be forced to use it

It's only the default formatting; you can reformat the cards if you like.

It's still evil on the part of Microsoft and incredibly stupid on the part of embedded device manufacturers (many of whom depend on Linux by now).

Re:Microsoft and Making Money

[Anonymous Coward]

Since exFAT apparently is referenced in the SD standard, people will be forced to use it, if they buy any consumer electronic device containing an SD slot. They can't choose not to use it. It's a hardware standard.

So after exFAT, they won't be able to do what they do today, that is, freely exchange their media among their devices at their will. That's evil, and once again, it comes from Microsoft.

Is there any reason why you can't use UDF on flash media? It's designed for media that wears out with too many writes, so it seems like a perfect fit. And recent (since ~2000) versions support Unicode, so you can use Tengwar Sindarin for your file names.

You might be able to. The camera or media player you buy can't.

To expand: The device manufacturers have to play nice both with the SD licensers and Microsoft. Plus they have to face the reality that independently fixing or replacing file system drivers on Windows (like adding UDF write support on XP) just won't fly.

Re:Microsoft and Making Money

[afidel]

CFast, the next generation Compact Flash spec does not specify a filesystem but the first devices to be announced have been FAT32, the big problem there is video files are limited to 2GB.

Re:Microsoft and Making Money

[afidel]

I think it mostly has to do with write support being very spotty across platforms (not that exFAT is any better in that regard obviously).

Re:Microsoft and Making Money

[hairy monster]

Or use http://www.ridgecrop.demon.co.uk/index.htm?guiformat.htm to format bigger than 32GB on XP and above.

Re:Microsoft and Making Money

[Anonymous Coward]

Plus they have to face the reality that independently fixing or replacing file system drivers on Windows (like adding UDF write support on XP) just won't fly.

Why not? XP won't support exFAT natively anyway, and CD burning software makers have done it for the purpose of using on rewritable CDs, so it is possible to do.

Latest in a long line of suck

[sco08y]

FAT looks like someone's half-baked science project. FAT32 and exFAT (aka FAT64) just take the same mistakes and repeat them.

The fact that FAT32 is widely available is irrelevant; everyone will still have to install drivers.

So, yes, there's a demand for a simple (needing little CPU and RAM) filesystem. There's even an argument to be made that it should honor the same overall contracts that FAT does so that device manufacturers don't have to put lots of extra logic in. But it does *not* need to be the spawn of FAT.

Re:Latest in a long line of suck

[dargaud]

FAT looks like someone's half-baked science project.

Quite so. I remember writing an experimental filesystem for 3" (not 3 1/2") floppies on the Oric in 1982, making up my own concepts as I had no experience in the matter. It didn't really work but it wsa good learning. Then a couple years later I looked at the details of FAT and was surprised by how simple, similar and limited it actually was.

Re:Latest in a long line of suck

[Megane]

Even the Macintosh started out with a FAT-like file system, called MFS. I've seen what it was like when you tried to use it on a hard drive with 5 megabyte partitions, and it was not pretty.

But Apple didn't just sit on it, and had HFS ready when they released the 800K floppy drive. HFS needed one major change (with MacOS 8.1) to support hard drives larger than 2 gigabytes (and more than 65536 allocation blocks per volume, also a big problem), and HFS+ has stuck around since then, though it has received a LOT of backward-compatible extensions in OS X, including journaling.

Re:Latest in a long line of suck

[yuhong]

Indeed the original version of FAT was I heard written by Bill Gates and Marc McDonald back in 1977. Tim Patterson extended it to 12-bit with it's SCP 86-DOS back in 1980, which was sold to Microsoft then licensed to IBM after it's talks with DR failed (another story altogether).

Re:Latest in a long line of suck

[yuhong]

But Apple didn't just sit on it, and had HFS ready when they released the 800K floppy drive.

To be more precise, when Apple released the HD20 which was a 20 MB hard drive for the Mac.

HFS needed one major change (with MacOS 8.1) to support hard drives larger than 2 gigabytes

Actually, large volume support was introduced with the PCI Power Macs and the PowerBook 5300 back in 1995. (If you read the developer notes on these machines, you will see that it has a chapter on large volume support.) It was ported back to all 68040 and PowerPC Macs with Mac OS 7.6. But large volume support with HFS was inefficient, as it was done by increasing the cluster size beyond 64 KB, making the problem of slack (which existed in both FAT16 and HFS) even worse.

Re:Latest in a long line of suck

[sco08y]

Yeah, FAT looked like a clever system when I was just learning and had never heard of a B+ tree.

From what I recall (FAT32 and exFAT both add structures and I'm only somewhat familiar with FAT32) FAT basically has a header (fixed at block 0, terrible for flash media) with a fixed list of root directory entries and the eponymous file allocation table. The FAT itself is just a list of every cluster in the media; each entry is either a link to the next block or an indicator if it's bad. Directories are allocated like files; they are a straight list of entries.

And that's *it*. Simplicity is fine, but, as common experience shows, it just isn't good enough.

ext is on MSWindows but not widely known

[NaCh0]

There are several quality open source programs I always use on Windows. Most know the big ones...Firefox and OpenOffice. Open Source advocates need to familiarize yourself with http://www.fs-driver.org. They have created an ext2 driver for windows. If this driver gains in popularity, it will be one less "Microsoft tax" you pay on your gadgetry.

Re:ext is on MSWindows but not widely known

[Sterling Christensen]

That driver has a serious user-unfriendly limitation: No support for inodes larger than 128 bytes.

This means Linux users can't use GUI tools to format a USB stick (or a harddisk partition for sharing files with Windows) - they must use the command line and figure out how to persuade mkfs.ext2 not to default to 256 byte inodes. And this probably after learning of this limitation the hard way. Easy enough for you and me, but definitely not user friendly.

Also, this still leaves Windows users unable to format as ext2. A crashy driver is not enough.

That brings me to the third problem: I have yet to see a stable IFS (Installable FileSystem) driver for Windows. In my experience, perfectly stable Windows installations start crashing when an IFS driver is installed and in use. I suspect this part of Windows needs more debugging, or the API needs to be better documented, or both.

exFAT may be a patent encumbered extension to a lame filesystem, but the ext2 drivers for Windows are a lousy counter proposal.

Re:ext is on MSWindows but not widely known

[jonbryce]

MacDrive seems stable enough, although it isn't free in any sense of the word.

Re:ext is on MSWindows but not widely known

[mrt_2394871]

That driver also needs to officially support Windows 7.

On RC1, I have to (re-)assign drive letters on every boot into Windows.

Re:ext is on MSWindows but not widely known

[JohnnyBGod]

Though it blue-screened almost constantly when I started using it, ext2fsd is pretty stable, these days.

Re:ext is on MSWindows but not widely known

[tepples]

They have created an ext2 driver for windows.

Not everybody owns the PC that he uses. Can this driver be used by someone who does not have privileges to install programs (user outside the Administrators group) or to run executables from %USERPROFILE% or removable media (look up Software Restriction Policy on a search engine)? I'm thinking of a scenario involving a user at a public library, Internet cafe, or employer's break room. Linux has the same thing: user not in sudoers and /home mounted noexec respectively.

Re:ext is on MSWindows but not widely known

[MobyDisk]

exFAT may be a patent encumbered extension to a lame filesystem, but the ext2 drivers for Windows are a lousy counter proposal.

Perhaps those specific drivers are. But the proposal of using ext2 makes perfect sense. It's open, unpatented, and already has free implementations. The only reason for manufacturers not to support it is fear of reprisal from Redmond.

Just great...

[MojoRilla]

And the best news...

The SD Association has adopted exFAT for its SDXC memory card specification.

So a mediocre but patent encumbered technology gets adopted as a standard because it runs out of the box on Windows. As Microsoft itself puts it, "exFAT is relatively simple". Hello, antitrust regulators? Hello, patent office?

Re:Just great...

[Anonymous Coward]

"mediocre"
Please justify your claim in the scope of intended use (portable media).
Otherwise it's just a troll.

And let me remind you, that ext2&3 do not support TB files without tweaking block size. And even after such tweaking SUSE 11 had huge problem with my single 2 TB file.

Re:Just great...

[abigsmurf]

Why is it anti-trust? SD cards, although popular do not represent a monopoly of the flash media market. History has shown that multiple file structures on portable media doesn't work as HW makers don't want to spend time and money including a dozen or so extra drivers on their devices.

Don't want to create a piece of hardware with support for exFAT? Use another memory format. No one is forcing you to use it.

Re:Just great...

[TheRaven64]

It's still a FAT variant, which means that seeking in a file is an O(n) operation (it's O(log(n)) on most systems) in terms of the size of the file. They've added a free space bitmap, but creating and appending to a file is still O(n) in terms of the size of the disk, just with a smaller constant. Both the FAT and the free space bitmap need to be kept in RAM for reasonable performance. The FAT size depends on the disk size and the configuration, but a typical 32GB memory card will need 32MB for the FAT. This is a lot of memory for a mobile device. Something like the N900 has 32GB of Flash and only 256MB of RAM. You're using an eighth of the RAM just for the FAT. More if you add another memory card, and that's not counting the free space bitmap (also needs to be in RAM, but is quite a bit smaller), ACL or file caches or any other driver overhead.

Oh, and the FAT itself needs to have individual words updated in a large contiguous section, which is about the slowest operation possible for Flash. They could improve this by using -1 instead of 0 to indicate free sectors: then allocating a sector would not require erasing a flash sector, but deallocating would.

Re:Just great...

[Anonymous Coward]

No one is forcing you to use it.

Really? So if I format the SD card with ext2 and shove it in my digital camera, it will successfully read and write photos to it, will it?

You keep using that phrase, I don't think it means what you think it means.

Re:Just great...

[abigsmurf]

Someone held a gun to your head and said 'you must buy a camera which only supports SD cards'? Don't blame Microsoft for something that is both your and the manufacturers choice. As I said, you're free to pick a camera that supports something other than SD, the camera manufacturers are free to make cameras that support other formats.

Re:Just great...

[Waffle Iron]

SD, the camera manufacturers are free to make cameras that support other formats.

But they won't, because Microsoft will use its privileged position as the sole controller of "security" updates for its desktop monopoly OS to automatically push this encumbered filesystem to the vast majority of computers in use. No camera or card maker could ever hope to surmount that barrier and install enough filesystem drivers to reach critical mass of general adoption.

Microsoft is leveraging its monopoly position OSes to generate royalties in the unrelated camera market. That looks like an antitrust violation to me.

Re:Just great...

[I_Wrote_This]

It's still a FAT variant, which means that seeking in a file is an O(n) operation (it's O(log(n)) on most systems)

Irrelevant for SDXC, though. I can't see that seeking.

(from earlier)ISO-9660 doesn't support Unicode.

UDF does. But MS only allows that on an optical drive. (Format a hard-drive with UDF and XP is flummoxed). MS == no imagination.

Re:Just great...

[abigsmurf]

Yeah, because nobody expects a file system to be part of an operating system!

Camera makers are free to use whatever file format they like. If it was one unsupported by Windows, simply include a file system driver to be installed alongside the drivers and utilities that come with most cameras.

Re:Just great...

[TheRaven64]

Irrelevant for SDXC, though. I can't see that seeking.

No it isn't. You need to read n entries in the fat to seek within a file. I'm talking about algorithmic efficiency of the data structures used, not about mechanical motion within a drive.

Re:Just great...

[Waffle Iron]

Yeah, because nobody expects a file system to be part of an operating system!

That's not the problem. The problem is that this royalty-generating file system will be part of most cameras and other gadgets.

If it was one unsupported by Windows, simply include a file system driver to be installed alongside the drivers and utilities that come with most cameras.

Countless others on this story have pointed out why that is not practical. Installing a driver takes significant effort and administrative rights on the system. User will not go through this effort if Microsoft has already installed their proprietary driver through their unique back channel that nobody else has access to.

In theory, nobody is "holding a gun their head" keep people from adopting a different file system for cameras. Also in theory, quantum fluctuations could cause a tiny flying pig to materialize on my desk. In reality, neither is going to happen. You know it, I know it, and Microsoft knows it.

Re:Just great...

[badkarmadayaccount]

I propose a specification that describes how to embed a CPU independent sandboxed fs driver in the superblock, and which is only writable by a special hardware device, not by software on a generic PC, thus mitigating infections.

SDHC readers

[backwardMechanic]

This is only tangentially related to TFA, but lets have a go:

I can't read SDHC cards on any Windows PC I can find. I have one of those cheap USB readers. It used to work, then it stopped - I have no idea what changed. It's not hardware, because Linux machines will read it just fine. Yes, the card is >4GB, but as I said, Linux can read it, so this is not a hardware problem. Does anyone out there know how to fix this?

Re:SDHC readers

[peppepz]

Microsoft realeased an updated SDHCI driver for windows XP which is supposed to support SDHC cards. But since it has not been distributed with Windows Update, I suppose it could have some problem. I've never tested it.

Re:SDHC readers

[rdebath]

Can't be sure, but the filesystem & partition driver for windows USB devices is sometimes very temperamental. Generally both the partition table and the filesystem must exactly match what windows expects.

If you're feeling brave you could try a complete wipe of the SD card partition table and all.
ie: "dd if=/dev/zero of=/dev/sdc"
Then reformat it under windows.

I would make sure I keep a copy of the existing partition table and filesystem though.

It also is supported by Windows XP

[hairyfeet]

You just have to apply the update, which you can get here. So has anybody benchmarked it to see how it compares to FAT?

And for those in Linux that want exFAT support according to the wiki an opensource experimental driver is in the works, or you can purchase a proprietary driver derived from licensed MSFT source code from Tuxera.

That said I doubt we will be seeing FAT go anywhere for awhile, even though FAT is pretty long in the tooth. Sadly FAT is the only format that I know of that can be truly read by all three OSes out of the box.

Re:It also is supported by Windows XP

[peppepz]

And for those in Linux that want exFAT support according to the wiki an opensource experimental driver is in the works, or you can purchase a proprietary driver derived from licensed MSFT source code from Tuxera.

The sad thing is, that exFAT is being patented. That means that Linux users will either have to be lucky enough to live in that shrinking part of the world where software patents are not allowed, or consult their lawyer before connecting their camera to their computer.

Oh well, at least the patent submission appears to contains a copy of the exFAT specification, so reverse engineering the format won't be that hard.

Re:It also is supported by Windows XP

[Anonymous Coward]

Or suck it and see. If this is an obvious extension of something done previously and microsoft want to get money for their patent, then they'll need to pursue that in court.

Re:It also is supported by Windows XP

[Tapewolf]

The sad thing is, that exFAT is being patented. That means that Linux users will either have to be lucky enough to live in that shrinking part of the world where software patents are not allowed, or consult their lawyer before connecting their camera to their computer.

That's the same situation we have now with NTFS-3g and portable disks, really. Doesn't seem to have stopped people. Though to be fair, Microsoft could well be more aggressive about licensing exFAT since no-one is going to want to implement NTFS on a camera anyway.

Re:It also is supported by Windows XP

[peppepz]

People can do without NTFS on a portable disk, because the only reason to use it is to interoperate with Windows. Which supports other file systems, so there is choice.
People won't be able to do without exFAT, because (if, and when, the standard gets adopted) it will be the file system used by consumer electronics devices. Which won't likely support more than one file system, so there will be no choice.

The beauty of digital storage for media is the freedom for the user to access his data in every way he sees fit. Closed standards for *personal* media storage are a step in the opposite direction.

Re:It also is supported by Windows XP

[hairyfeet]

Then why isn't the FLOSS guys pushing an alternative? It seems like it wouldn't be too difficult to either have a "FAT translation layer" that converts to whatever FLOSS file system is on the device, or perhaps just have Win98-Win7 drivers stowed in a small FAT read only partition on the device?

It seems like if the Linux guys want to be able to use these devices in the future NOW is the time to be coming up with a non patented alternative. Last I checked EXT3 drivers for windows were kinda buggy, bad enough I probably wouldn't want to trust my data to them, and while NTFS works on Linux with NTFS-3G (IIRC) NTFS really isn't the right format for flash devices. So what alternative is there? FAT32 is also patented is getting too long in the tooth to handle the exploding media file sizes. So unless something else comes along it looks like exFAT will be the next format for flash based devices.

"The technology"

[Lord Lode]

"The technology is available on Windows 7, Vista SP1, Windows Server 2008 and Windows Embedded CE".
I find this as technological as a fork or a pencil are "a technology". Why do trivial things so often get called "the technology"?

Re:"The technology"

[MrMr]

It's relative to the user. It's just like "a challenge", when you're an incompetent idiot everything looks like one.

Sounds like a weight loss program infomercial

[cheap.computer]

EX-FAT sounds like some kinda weight loss program. Or is msft really losing its weight in mobile space ??

this is an ad!

[mxh83]

"Microsoft hopes companies making devices such as cameras and smartphones will adopt the Extended File Allocation Table (exFAT) technology to support the sharing of audio and video files. The technology is available on Windows 7, Vista SP1, Windows Server 2008 and Windows Embedded CE." Ok, so why are you advertising it here? How do such topics make it to the front page?

Re:this is an ad!

[Megane]

Sounds more like a warning to me.

already adopted and also for XP

[Nabbler]

exFAT was adopted by the SD group as a standard for their super-high capacity SD cards, so that's covered already. And also it's available as an update for Windows XP too.

ext2 on windows.

[Anonymous Coward]

If only there were ext2 drivers for Windows . . .

Oh, wait.

(I personally recommend ext2fsd of those listed)

FAT patent expiry

[Alain Williams]

When does the FAT patent expire ? Getting manufacturers to adopt a new, patent encumbered, standard will give them another 20 years of harassing Linux users.

Hey Bill ... err ... Steve ... just make it free

[Skapare]

... and they will come!

No Mac? No, thanks.

[vrmlguy]

Seriously, until and unless there's (at a minimum) royalty-free supported drivers for non-Microsoft and pre-Vista (not just XP, I use 98 in some situations) operating systems, I don't see this going anywhere.

MTP

[DrXym]

If I were developing a multimedia box, or an essentially closed system (e.g. a satnav), what possible reason have I got for using exFAT? The reality is I could just shove ext2 or any other random fs on my device and it doesn't matter.

And if its a device that plugs into a PC, why not use MTP? After all, MTP is fairly decent these days and it means that USB stick, or media player or whatever you plug into your computer (of any kind), can implement any fs its little heart desires. It doesn't matter because it is under the covers. If the device wanted it could even store files with FAT32, using MTP to break them up into chunks as they were read or written.

Perhaps MTP is too much for some simpler devices. It isn't beyond the bounds of reason to envisage a simple standard for multi-part files, (e.g. a manifest + file chunks) that would allow large files to be copied to and from even FAT32 devices. Linux could even take the lead here by implementing that support in its FAT32 device driver. i.e, that if you copy a > 4Gb file to a FAT32 partition, that it is physically split up into chunks but appears as a single file to the caller.

sharing of audio and video files

[Ozric]

We all know it's nice to share, or that is the way I was raised, but the last time I checked that kind of activity could land your ass in JAIL.

Kinda of a mix message, No?

One step vs. three: convincing administrators

[tepples]

If you use a widely supported file system on your camera, you have to develop a file system driver only for your camera. But if you use a Free file system on your camera, you have to develop a file system driver for your camera, develop a file system driver for Windows and Mac OS X, and convince PC administrators to install this file system driver for Windows and Mac OS X.

Re:One step vs. three: convincing administrators

[donaldm]

The ext3 file-system is widely supported on all Linux systems (well over 100 million machines) and it is free to use by any manufacturer so there is little if any development costs. Any USB or Memory card can be formatted with an ext2, ext3 or even ext4 file-system from just about any Linux distribution. The mke2fs (there are others) command that can do this is only 65kB in size.

Even if you need to transfer data from an ext3 file-system to an MS Windows machine you can always get software that can read that file-system. Here is a MS Windows ext2/3 reader if you don't believe me. Even Mac's have software that can read and format ext3 file-systems. So were are the development costs since the products are already available and are, shock/horror free?

Convincing PC administrators to install the appropriate software to read say an ext3 file-system is easy, it would go something like this. "We require software that can read ext3 file-systems. Here is a change request that is signed by the appropriate people to install the appropriate software". This normally works in the corporate world and it is not a problem for a vendor of a product that uses an ext3 file-system to provide the reader for the average PC home user. They do that now anyway for printers and many other peripherals so the same can be done for storage devices.

The only thing holding back manufacturers from using unencumbered file-systems are Microsoft apologists. It really is quite sad when you think about it.

Re:One step vs. three: convincing administrators

[tepples]

Here is a MS Windows ext2/3 reader if you don't believe me.

Does it run on a user account that is not a member of the Administrators group? And it appears to be read-only; am I going to have to carry a separate, smaller stick formatted in FAT32 for files that I want to copy to a Linux system?

We require software that can read ext3 file-systems. Here is a change request that is signed by the appropriate people to install the appropriate software

You mean a petition? Please. Do operators of PCs in public libraries, Internet cafes, and office break rooms actually read those?

Re:One step vs. three: convincing administrators

[u.hertlein]

Even if you need to transfer data from an ext3 file-system to an MS Windows machine you can always get software that can read that file-system. Here is a MS Windows ext2/3 reader if you don't believe me. Even Mac's have software that can read and format ext3 file-systems. So were are the development costs since the products are already available and are, shock/horror free?

As much as I would like to have a universal non-proprietary file system this won't fly with users. Both options are read-only which means you can't even remove images from the media after you've copied it to your system.

AFAICS the Win solution that you ref'd is a standalone program, meaning it's not possible to use standard file I/O calls to read from the device. So the pipeline would be:

• start reader
• copy to local system
• remove copied files on the camera (which ones did I copy again?)
• start GIMP, and get to work...

instead of

• start GIMP, load file, and get to work...

Re:One step vs. three: convincing administrators

[SiChemist]

So, you want an installable file system. You mean like this?

Ext2 Installable File System For Windows

I'm not sure why the GP referred to the sysinternals utility when there is a freeware installable filesystem driver available.

Re:One step vs. three: convincing administrators

[tepples]

So, you want an installable file system.

And privileges to install it. I don't find it likely that change requests to install Ext2 IFS on publicly available computers will be accepted more often than not.

UTF-8

[tepples]

Also, does it support Unicode?

Anything that supports the full 8-bit range of code units supports Unicode in the UTF-8 encoding.

Re:UTF-8

[greed]

Microsoft hates UTF-8, which was designed carefully so it wouldn't be damaged by existing 8-bit clean code. Microsoft actually uses UCS-16LE as an on-disk and on-wire format (not just an in-memory format).

This isn't the only case where they write out in-memory structures directly to disk; if you've been following the OOXML debacle at all, it's littered with in-memory blogs with XML decoration.

Re:UTF-8

[shutdown -p now]

Anything that supports the full 8-bit range of code units supports Unicode in the UTF-8 encoding.

In theory. In practice, when we speak of some format "supporting Unicode", what we mean is that either the format mandates UTF-8, or vast majority of implementations supporting this format read and write in UTF-8 by convention.

If you have a single application that writes in UTF-8, and everyone else is going to treat it as, say, Latin-1, then it's not really "support".

Re:UTF-8

[spitzak]

In Windows the file system api is 16 bits, so the driver would translate the UTF-8 to the 16-bit filenames and no program could get it "wrong".

If an 8-bit api like on Unix is used, you are seriously wrong about the problems with programs treating UTF-8 as ISO8859-1. They will display gibberish, yes, but they will not destroy the UTF-8 when copying the files and will not misinterpret it (as all the bytes with the high bit set have no meaning to the file system apis). The only failure is in display. Claims that somehow the text is "wrong" when stored in a byte stream are bogus, it is like saying that we must throw exceptions when misspelled words are encountered in text files. This would obviously make text handling impossible, but people try to push these ideas into UTF-8 because they want to sabotage it and preserve their incorrect but politcially-correct decision to use "wide characters".

Re:UTF-8

[shutdown -p now]

If an 8-bit api like on Unix is used, you are seriously wrong about the problems with programs treating UTF-8 as ISO8859-1. They will display gibberish, yes, but they will not destroy the UTF-8 when copying the files and will not misinterpret it (as all the bytes with the high bit set have no meaning to the file system apis). The only failure is in display.

So long as it's strings, a lot of operations that are string-specific will give wrong results. Examples are lowercasing and uppercasing.

And note that we speak about file names here, not file content.

In any case, I'm not talking about this from dev perspective, I'm talking about user perspective. Yes, any 8-bit clean encoding is theoretically capable of storing UTF-8. But if no application that works with that will properly display it and manipulate it, then, as a user, I do not have Unicode support - it's as simple as that. I don't care about it being "preserved" and other stuff.

Oh, and screw UTF-16, too. The only proper Unicode encoding is UCS4; everything else is a hack of various proportions (though granted, UTF-16 is a bigger hack than UTF-8).

Re:UTF-8

[spitzak]

Even with UCS4 you have combining characters, directional and spacing marks, and all kinds of other cruft. You have to realize the idea of thinking about splitting text into fixed-sized units that are independent of each other and can be rearranged is long obsolete. For this reason I think UTF-8 is far superior because it FORCES the lazy programmers to realize what is going on, rather than letting them pretend that only weird backwards countries are going to be inconvienced by their mistakes.

Also UTF-8 survives case shifting of ASCII. If you try to case-shift the ISO-8859-1 characters you will always end up with invalid UTF-8 (because the lower and upper ISO-8859-1 characters are the leading and continuation UTF-8 sets) and in fact this is fairly easy to detect and reverse. UTF-8 is very redundant, invalid UTF-8 sequences of 4 bytes outnumber valid ones by more than 10 times, making it extremely unlikely that any mistaken processing cannot be recognized and remedied.

Re:UTF-8

[shutdown -p now]

Even with UCS4 you have combining characters, directional and spacing marks, and all kinds of other cruft. You have to realize the idea of thinking about splitting text into fixed-sized units that are independent of each other and can be rearranged is long obsolete. For this reason I think UTF-8 is far superior because it FORCES the lazy programmers to realize what is going on, rather than letting them pretend that only weird backwards countries are going to be inconvienced by their mistakes.

I understand the issues that you mention, but I don't think that UTF-8 is any more helpful in mitigating them. All too often I see Western (and particularly Anglophone) programmers assume 1 byte = 1 char in UTF-8, simply because it works for any text that they deal with. It could be argued that with UCS4, more programmers have such an "excuse" of being lazy, but the problem remains the same. You can't really do anything about it apart from educating people.

The problem with using UTF-8 (or any other encoding that permits single byte per char) strings, as I see it, is that it encourages people to use the same containers for both raw binary and textual data, even though semantics differ. For example, in C++, every now and then you see std::string being used for binary data, even though std::vector<char> should be preferable; ditto for Dephi and AnsiString. On the other hand, this is much rarer in Java and .NET, where strings are UTF-16 (though, judging by StackOverflow, people still do occasionally try to do that) - since string doesn't look like an "array of bytes" anymore, one is less enticed to use it as such.

Similarly in Python 2.x and Ruby, people very often code against multibyte strings, and assume 1 byte = 1 char, but this is much rarer with wide strings.

Re:UTF-8

[hitmark]

hell, the binary ms office file formats are basically memory dumps of the state of the document at the time of the save...

You are breaking the DMCA ;)

[Ilgaz]

You know, MS (and various closed source) things come with "You may not disassemble the product". I am trying to show what is wrong with using a frankestein filesystem which should be ditched the _day_ floppy diskettes were over.

What a sad thing that Linux, BSD and even Apple couldn't convince these SD card idiots even after the Tomtom wake up call. If it is the popularity, iPhone/iPod uses HFS+ journaled for years with gigantic files in them. Near all "media device" powered by Linux runs some sort of ext2/3.

I would personally trade NTFS with FAT, at least it is modern and journaled. BTW, am I reading wrong or Nokia still uses God damn FAT in a Linux powered device? They are beyond hopeless. If anyone talks about "but they come with FAT", please don't. A Nokia phone does a lot of things with a freshly installed memory card (creates own dirs), how hard would be to check if it is empty and format it with ext2? "Easily remove" was _never_ the case, all Nokia devices have "remove memory card" option for a very good reason.

What bothers me is, non technical people store impossible to reproduce memories (photos etc) to that backwards filesystem thanks to these SD card idiots. We will see their faces when MS becomes a patent troll or decides to punish some company for using Linux.

Re:You are breaking the DMCA ;)

[TheRaven64]

BTW, am I reading wrong or Nokia still uses God damn FAT in a Linux powered device?

Only on removable flash cards. They use JFFS2 on the internal flash drives for Maemo devices, but FAT for removable media: people complain when they pull the flash card from their device and can't read it on a Mac or Windows PC (or even a FreeBSD machine if it uses JFFS2; it's Linux-only). The same with the phones. I transfer small files to and from my phone with Bluetooth, but it's much faster to just pop the card out and pop it in a card reader for large files. If it used JFFS2 (or ext2fs) then I wouldn't be able to read it in my Mac.

Could any MS file system be that impressive?

[tjstork]

They are going to try and foist on the world a gussied up FAT, and the question really is, when there's a small army of uber geeks writing free file and open source file systems for Linux, could a gussied up FAT be worth actually licensing? That's one thing that actually sucked me into Linux, is that, you have -so many- choices of file systems available, and these days if it has a file on a device of some kind, Linux can read it.

I don't mean to knock the possibility that some software can command a premium of paying for it, but, I think by its very nature - the smallish scope of the project, the well defined goals, and the diverse range of existing choices... do you really need to pay for a file system? I mean, NTFS is Microsoft's crown jewel file system, and based on my own benchmarking, and gut feel, I think EXT3 is faster, and its free, and if Microsoft can't beat that, why would I even consider a FAT variant, which is worse than NTFS by Microsoft's own admission.

Re:Could any MS file system be that impressive?

[Bert64]

Your point illustrates exactly why microsoft having so much influence over the industry is such a bad thing, and why many people despise the way microsoft do business...

Because of their size and influence, the world will end up stuck with the inferior exfat filesystem regardless of what else is available or how superior it is... MS will achieve this by ensuring their widely used os simply doesn't support anything else out of the box, making exfat the only option for many... This is also how fat32 got so widespread, despite also being total garbage.

Re:Could any MS file system be that impressive?

[shutdown -p now]

I mean, NTFS is Microsoft's crown jewel file system, and based on my own benchmarking, and gut feel, I think EXT3 is faster

There's more to it than just speed (and I've seen benchmarks go either way between ext3 and NTFS). Have a look at the feature matrix, as well.

Looks like another attempt to bump off Win XP...

[Anonymous Coward]

What better way for Microsoft to kill Windows XP than to make new portable multimedia players artificially incompatible with it? Here's a stroke of genius... PMP makers should use a file system from a third-party and include an installable file system driver for the PC. This way, Microsoft can't leverage the PMP market to attempt to force people to upgrade to Vista or 7. As a side affect, the PMP maker wouldn't have to pay a bullshit filesystem tax for using antiquated crap like FAT32.

Letting the company with the monopoly dictate standards for another market is a *BAD* idea.

UTF-16 vs. autoSpaceLikeWord95

[tepples]

Microsoft actually uses UCS-16LE as an on-disk and on-wire format (not just an in-memory format).

True, Microsoft has a hard-on for little-endian UTF-16. But at least this encoding is standard, and it's no problem as long as either

• A. the use of UTF-16 is documented elsewhere (e.g. in a file format spec or in protocol metadata),
• B. the use of Unicode in an unspecified encoding is documented and strings have a byte order mark, or
• C. the use of Unicode in an unspecified encoding is documented and there are enough Latin-1 characters at the start that the heuristic can reliably determine that the encoding is UTF-16.

But you're right about things like "autoSpaceLikeWord95".

"benchmarks"

[peppepz]

So has anybody benchmarked it to see how it compares to FAT?

I've done a poor man's benchmark by untarring the gnu glibc 2.11 tarball onto a newly formatted 2 GB usb stick under Windows. The stick was formatted with the default options for each filesystem, and was "safely removed" and reinserted between each step.

Speed: time it took to extract the tarball on the newly formatted device.
1) UDF 2m 33.171s, 2) NTFS 25m 29.718s, 3) exFAT 42m 30.390s, 4) FAT32 47m 31.640s.

Speed: time it took to delete the whole tree created at the previous step.
1) UDF 0m 34.875s, 2) NTFS 6m 28.156s, 3) FAT32 16m 10.578s, 4) exFAT 28m 18.000s

Disk usage overhead: space that was nominally reported as free just after formatting.
1) exFAT 2,047,410,176 bytes, 2) UDF 2,047,056,384 bytes, 3) FAT32 2,043,637,760 bytes, 4) NTFS 2,008,457,216 bytes

Disk usage overhead: space that was nominally reported as free ater extracting the tarball.
1) UDF 1,941,765,632 bytes, 2) FAT32 1,915,809,792 bytes, 3) NTFS 1,877,790,720 bytes, 4) exFAT 1,616,379,904 bytes.

Bad excuse.

[DrYak]

The SDXC committee needed the memory cards to be usable on Windows, ideally without installing any drivers, and without having to screw around to get it to work. {...} That leaves UDF, which can not be used on anything but optical discs in Windows XP, {...} FAT32 can actually scale well into the terabyte range, but Windows will refuse to format a disk as FAT32 if it's larger than 32GB. {...} Microsoft's solution was exFAT. The selection of exFAT instead of FAT32 was forced by limitations on Windows' FAT32 support.

Well, exFAT isn't supported out-of-the-box in Windows XP either. And if you're going to accept exFAT, then you accept a system which only supported out-of-the-box in modern versions and requires an update installation on older versions.

- UDF pretty much follows the exact same description. Its works in read-write mode on modern Windowses. And under XP it requires a (3rd party) update to get write capability: It's even better than exFAT given the fact that, prior installation, at least WinXP has read capabilities. (Enabling to make a prompt for updating WinXP as an autorun upon media insertion. Not very secure, but that's what windows users are used too anyway)

- FAT32, the point is moot : nearly 100% of media is currently sold pre-formatted and even older version of Windows are able to use it. Windows will only refuse to *format* it, not to use it if pre-formatted. User seldom format media themselves as its pre-formatted and as their preferred method of erasing is drag-dropping stuff into trash.
And all it takes is an update to enable the formatting tools to handle such bigger partitions.

So in short, the other two solution have similar advantage and problems as exFAT regarding availability on Windows *AND* happen to be available on other platforms such as Mac OS X (even if you are going to ignore Linux under the pretext it's less popular). In fact UDF is also available on lots of embed system, given that it's necessary to play optical media. So people who just plug their USB stick or SD card into the (already connected) home entertainment system (DVD, Blueray, PS3, whatever) instead of bringing the laptop to the TV-screen or finding the cable to connect the camera to the TV screen, could soon do it with UDF media (it's only a small modification to the firmware to have it expect UDF as a possible filesystem on a media - whereas exFAT would require licensing and developing a whole new file system driver much more difficult until new device / firmware updates hit the market).

The others

[DrYak]

Are you telling me that it's impossible for me to format an SD card in something else than exFAT ?

No. But the future SDXC cards will be sold pre-formated on exFAT (instead of FAT32), and the "erase card" options of future SDXC-compatible cameras is going to use it by default too.

Thus, although you, as a knowledgeable geek, will probably manually format your SD media using FAT32 (for anything that goes in digital cameras) and UDF (for anything which only goes into Linux, Mac OS X and numerous Multimedia Players / Harddisk enclosures and media centers (of which, lots happen to run Linux too) ), most of the media circulating "in the wild" will be formated in exFAT, for example SDXC card that non geek friends could give to you.

And although *you* will personally will be probably able to use as self-compiled version of the (currently still experimental) exFAT Linux drivers, any commercial manufacturer attempting to implement it into their gadgets or any official developer including it into an official Linux distribution is going to get sued-smashed "Tom-Tom & vfat" style. (Although you'll probably find it in Debian-nonUS)

FAT should die, should really die

[Ilgaz]

Issue is, Nokia already installs gigantic amount of device drivers, frameworks to the host OS (Windows) and it would be trivial to add a ext2 vfs to the set. Nobody would bother either, at least the people who thinks having 290 MB of application suite (PC Suite) to communicate with their device is normal.

I am not saying Apple does some sort of magic thing either, their support files are huge, equal amount of "risky", "deep running" device drivers installed.

Wonder what if the general public knows how risky and dangerous running FAT is, especially with removable, gigantic, high end memory cards and offered an easy to install (just like Adobe Flash), tiny "driver"? At last resort, they could do the trick Apple did in HFS+ transtition. Put the driver to the first partition which will also be locked, put some informative readme files (in all languages) and a decent, WHQL certified driver which will be digitally signed and comes in MSI/PKG files for both Windows and OS X?

It is fine up to 2 GB but really, having 32GB cards running that backwards filesystem has become the IT comedy of all times. Even MS does every trick on UI book to "save you" from formatting a drive in FAT these days.

Funny enough, HPFS (grandfather of NTFS) is a full MS technology. They knew FAT is going nowhere back in 1990s so the first thing they did was getting rid of it... At least until the fight broke with IBM. I believe the FS of Xenix (MS Unix, older than MS-DOS) was more advanced than FAT already. Just like they had to copy CP/M, they had to copy its backwards filesystem too. You know, IBM even asked for a CP/M like prompt.

If I definitely need Windows compatibility, I even format the 2 GB USB sticks in NTFS format. At least I know its roots and how it operates. I stay away from FAT as long as possible.

Re:FAT should die, should really die

[Jesus_666]

The problem is that Microsoft doesn't really support third-party IFS drivers. They have APIs for that but they don't offer any help and as a result most IFSes (includeing the two available for ext2) are nowhere near the quality you'd expect from a file system driver. I'd much rather trust my data to FAT32 than to an ext2 partition regularly written to by Windows.

Oh, and IIRC the ext2 IFSes don't support fsck; you have to do that on a Linux box. And they don't support block sizes >128 KB so you have to use non-default parameters when formatting, which the IFSes also don't support.

Of course Nokia could theoretically write their own Windows implementation of ext2 from scratch - but the result would be that now they have another Windows driver and a handful of userland programs to maintain and their customer are stuck with a file system that virtually ensures they can never put the data on a computer that doesn't run their software. In short, a customer service nightmare.

Or they just use FAT32 and their customers silently accept that you just can't store very large files on flash cards.