Windows XP driver to read ext3 partitions?

http://www.fs-driver.org/

That seems to be able the only solution. All links point to there.

Ext3 support in other OSes is pretty dang good due to it being an open FS.

In as far as NTFS solutions for linux, there are a few, the captive driver solution probably being the best free (As in beer due to closed source code being used) solution, however there are a few commercial solutions availiable I haven't looked into in a LONG time.

Seablade

The problem with going to a Linux FS for long term storage is that you become dependent on the OSS community to develop drivers for future Win FS's. Either that, or use Linux to transfer your stored files to an older Win FS and then use Windows to transfer to the new Win FS. NT was not a consumer OS, but both XP and Vista still support FAT, and follow-on Win OS's will likely continue to do so in part because FAT formats are still popular on flash media and devices.

You might want to reconsider going to NTFS. Even though the in-kernel driver is marked "Experimental", support for NTFS is considered by many to be stable for simple file transfers. If you're really concerned, or are planning to use it extensively, there is an external module called NTFS-3G (ntfs-3g.org) that many people use. NTFS-3G is also considered stable (and actually better than Windows NTFS in some ways) even though it too is claimed to be beta. I don't think MS will be doing anything with the NTFS format that would make it incompatible with Linux drivers, but I would format your storage partition using Windows just in case.

I believe you meant Win OS's, but that aside, why is that any worse than depending on closed source companies to provide future drivers for the same thing? Seablade

I was attempting to project 20+ years or so into the future (what I think of as "long term"). I have more faith that MS will maintain backwards compatibility with NTFS than I do in the OSS community to develop drivers for a future Windows OS for what will likely be at that time an obsolete Linux FS. At the other end of the problem, I also believe that OSS devs will have a difficult time writing drivers for interoperability with Windows FS's in the future as those FS's start to resemble Windows proprietary relational DB's.

I completely disagree with your reasoning. Actually between the two I would be more confident of the OSS developers maintaining a useable solution myself than a closed source company that deems it unprofitable. Aside from that, you are suggesting using a reverse engineered driver that is not confirmed to be stable(Though many do use it admitedly) over one that is very strongly encouraged to be used by the kernel folks in charge of it? Between the two options I would say ext is the MUCH safer bet, it does not have to be reverse engineered, so less chance of failure in that regards, and is being fairly well supported on all platforms right now. Seablade

The OP stated that he wanted to be able to access his data from Windows which means a Windows compatible FS. I think it's safe to assume that Windows will still be the dominant consumer OS 20 years from now, I really don't know what will happen to Linux.

It seems to me that your fears about MS dropping support for legacy filesystems don't hold weight through either historical experience or logical projection. NTFS has a huge installed base, as it is now about 15 years old, and is the primary FS for Vista which will likely be MS's OS for at least 5 more years. Beyond the fact that NTFS is the OS FS of choice, it is also poised to become the storage standard of choice for flash devices and portable electronics that likely wont be maintained as well as MS's OSes. So if Vista has native support for Windows filesystems that were used 20 years ago in the first Windows OS version (when portable storage was done via floppy drives), then I'd say that the chances are very good they will be supporting NTFS in Windows 20 years from now.

Its hard to quote histrorical reasononing for 20 years in computer time, as PCs are still not much older than 20 years. And I never assume any company will be around for all 20 years, that is a mistake in itself. But at any rate if you really want historical reasononing, consider that Linux has been around nearly those 20 years as well Its not exactly a newcomer either. Open Source Software has been around longer, and still has support for those filesystems as well. I don't think there is any particular reasoning to choose closed source over open source there, and it is easier down the road to. worst case sceanrio, get a FS driver written for an open FS than a closed one.

In as far as maintaining backwards compatibility, MS has started to learn its lesson with that, thus why you don't have backwards compatibility with all old software in Windows.

And NTFS DOES have a decent installed base, but if you think that it is going to remain MS's FS of choice over the next 20 years I have a bridge to sell you. NTFS is dated and in terms of FSs it is starting to show its age. I would strongly suspect them to replace it before to long myself.

Seablade

And since for some reason I can't edit my last post, I meant to point out the OP wanted a solution that can access files from Windows AND Linux.

Seablade

IMHO any OS tied to the GPL will always be a niche product at best. Even if Linux were technically competitive with MS (and it's still not after 20 years of development), there are just too many hurdles to overcome. Without commitments to established standards (impossible under the GPL I think), Linux is both too risky and expensive for commercial vendors to develop for. MS *is* a standard on its own, and well entrenched in the market. Consider that companies like Ford, AT&T, and IBM also became huge monopolies in their markets (and they had competition); you know how old those companies are... In fact one could argue that there might be compelling public interest for goverment subsidization if it ever were to be needed.

As I've stated previously, I expect MS's next gen FS win be incorporated into a future Vista release. However, for the reasons given, I still expect NTFS to be supported well into the future. Also, given NTFS's installed base, I expect there will always be strong demand for Linux NTFS support. So even if NTFS support is dropped in future kernels (which I don't think is likely to happen), I expect there will still be 3rd party developers writing NTFS compatibility drivers for it just as there are now.

Im not so sure if anyone will write a Windows compatibility driver for an old Linux FS and a future Windows OS. Moreover, I expect that with his files on NTFS, those files will be easily accessable throughout their lifetime using consumer devices that interoperate with Windows. Using a Linux FS would require him to do some conversion work, and everyone hates that don't they...

Starcub, are you trying to start a flamewar here? Nobody cares that you think MS is great and that linux is a little niche hacker's OS. The purpose of this thread is to discuss file systems compatible with windows and linux.

If you want to start an MS vs. Linux thread, feel free to do so, but make sure you are respectful and back up your points with references. Honestly, nobody gives a crap about your hunches or intuition. Present some well thought out opinions with references and you'll start a productive discussion. Saying things like Linux is not technically competitive with MS after 20 years of trying just makes you sound ignorant.

And?


LAMP. Just because it doesn't compete in all markets doesn't mean it doesn't compete. BSD is the same way. *nix based systems are known to be not uncommon in enterprise server environments.


Exactly how are there LESS hurdles with developing a driver to an OPEN file system, than developing a driver to something that has to be reverse engineered?


Again false. See SAMBA.


See above about enterprise.

Aside from that, Linux being risky or expensive(Another assumption as I doubt very much it is any more expensive to develop for than any other OS) has nothing to do with commercial vendors developing for it. That has entirely to do with ROI, which does contain expense in it yes. However the development expense has much less to do with it at that point compared to whether they would profit from the userbase on it. In other words, that all comes back to the userbase and the size of it.


Ford became a Monopoly, and no longer is.
AT&T got broken up.
IBM used to be a monopoly in computers in fact, no longer is. IBM-DOS anyone? OS/2?

Monopoly does not mean permanant. Apple has been around for a good long while as well, but they are hardly a monopoly.


And this means ext won't be?


Well to be honest, noone is going to write a driver for a FILESYSTEM for compatibility because they are locked to that Filesystem unless they are using a specific OS, in which case they don't need to write that driver, as it is generally already in the OS. Networking does not use the FS driver except on the server, so that is out as well. I can't think of a single reason to do that. Maybe, just maybe if in the future you are dependant on a machine running Win2k or some other NTFS dependant computer, and need to transfer files with permission data between them, but dang that is one heck of a stretch, there are to many other options/standards out there for you to use(iso, FAT, via network, etc)

When people write drivers because they are forced to for compatibility reasons, it is because they depend on certain HARDWARE. Whether that be measurement hardware specific to the scientific community(The most common occurance) or just old hardware someone wants to play with, like say the Midi Synth out of a Nintendo(Less common, hobbyist market, but still happens). But all those are directly tied to the hardware. The filesystem is not tied to the hardware. If you need compatibility for the Filesystem it is for specific reasons, all of which involve a machine physically running an OS that depends on NTFS and communicating with another, out of which, virtualization will be the norm in 5 years and right now that uses network processes to communicate with other filesystems.



Um, the OP was discussing Dual-Boot, not a removeable device. Thus that entire last point there is VERY moot, as not consumer device directly accesses the filesystem without going through the OS unless it is directly connected to the HD itself at the moment.

Now all this aside, what happens 20 years down the road, will make absolutely crap of difference to the OP because 5 years down the road they will most likely be getting a new computer again, and that old HD will likely be either reformatted or dumped. If data is needed off it, it will be transfered first, and thus be on a new FS(Possibly) and it makes absolutely no difference.

What it really comes down to for the OP, is which do they want to do, install a driver that had to be reverse engineered on Linux, or install a driver for an open filesystem on Windows. Between the two chances are the open filesystem is going to be the much better choice for a dual boot system on the same HD that will only be seen by two instances of any OS.

Seablade

Actually I was hoping to avoid an MS vs Linux debate, as IMHO the debate on that question was concluded long ago, and there's no new data to change the conclusion. My rationale for claiming that NTFS would be persisently supported was called into question and I felt it reasonable to establish marketshare and longevity as supportive of my opinion.

Have I said something disrespectful or illogical? I try not to, but people define respect differently and you never know how someone might take what you write on the net.

As an example, this statement of yours is one I personnally consider disrespectful. Why? Because "crap" is a slang term with negative connotations. A better alternative would have been to limit your complaint to specific issues and discuss what you didn't agree with about those particular issues. If you do that, then you might earn my respect and you would stand a better chance at convincing me of your opinions. Otherwise, that negative slang makes it difficult to dialog in good faith. Anger begets anger, and that is definately an impediment to constructive dialog. I tend to turn people off if they demonstrate disrespect, even if they have something valuable to add to the debate (like Seablade's reference to SAMBA which I hadn't considered previously). Like most people, I am a creature of emotion, and at times that can be a fault.

Sorry if I came across as belittling Linux, that wasn't my intention at all. I really don't like MS all that much because of the anti-competitive business practices they engage in. However, I don't think there is any debating the fact that MS produces a superior product for a large userbase. It's a highly integrated OS that is easier than Linux to install, configure, and use for the average user. In addition you don't have to worry about getting your new hardware to work with Windows. I think most objective users of both products would conceed those points along with the point that Linux serves a niche market (the technically inclined). Consider that Linux has been free for a long time now and still has a small share of the PC market, not for artificial market reasons, but for reasons I believe are endemic to large scale OSS -- that's broadly applicabale, again I'm not trying to belittle Linux (which I'm glad exists and use occasionally myself).

The fact that Linux is tied to the GPL IMHO means that standards will always be elusive. OSS filesystems, x-server, desktop gui, hardware drivers, etc... these all change and new projects are created and die, or fork according to developer interest. In fact it's concieveable that some OSS *nix derivative could displace Linux as the top OSS OS 20 years from now; how well would Linux be maintained then? A small risk perhaps, but still... I just think that MS is the better solution and probably will be for the next 20 years.

Well, Starcub, sorry for saying you're opinion is crap. The problem is that you say things like "I don't there is any debating MS produces a superior product."

Have you actually tried to install a modern distro? Have you ever used a live disc? Have you ever tried to install windows on an sata raid controller? The install process on winxp is not simple, we're all simply used to it. The only thing easier than installing OS via a live disc is getting it preinstalled from the manufacturer. Reality is most people haven't a clue how to install windows.

Once windows is installed you still need a driver disc or hit the net to find drivers. Linux, on the other hand, has virtually all of the drivers included. Of the 4 linux system I have all of their hardware is supported perfectly with no driver issues.

I find it funny that your criticisms of linux don't even touch on the real problems of OSS and linux in general. Most of your crits are considered positives by most of us.

Anyway, I don't have a few hours to disect your post, but I don't think you actually understand the purpose of OSS. You want linux to be like windows in more ways than one... which is the wrong perspective and you will always be disappointed if you think that way.

Now this is new to me. Couldn't find it in wiki...

True, but my intention was to show why I didn't think Linux would never obtain a foothold in the consumer market. In the consumer market, widespread adoption of any product (especially where there is no serious competition) in the marketplace tends to secure both the longevity of a company, and the degree to which a product is supported.

My question is who is going to do it if that Linux FS becomes outdated? I have a high degree of confidence based on both past performance and future useage predictions (reasons for both I've already given) that NTFS will be supported well into the future. Is there an analog for ext3? What FS did the first Linux release use, and is there still a driver for it in today's kernel?

SAMBA development was driven by an industry need to exchange data between Unix and Windows networks (filesystem independent). There was broad compelling interest that had alot of money behind it, and I expect that need will be persistenent as long as the two different OS's exist. I would not put filesystem support (a continuously developing tech that independant dev's maintain themselves) in that class.

You outlined how it does: expense impacts ROI (through development costs), as does the size of the userbase.

If I were developing a Linux program that I wanted to sell on the commercial market, I would want for that program to run on the largest possible userbase with a minimum of effort. To sell to a large userbase, I would need to write for several kernel versions, several x-servers, several window mangers, several filesystem structure possibilities, interface with different hardware driver possibilities, etc... that cost adds up. This is probably a large part of why you can't find many Linux products in the commercial market. So Linux (along with every other popular OSS OS) is almost entirely dependent on OSS developers to make it work.

Yes, but those very old companies are all still around due in large part to fact that they were monopoly's. When a company becomes a monopoly on the free market, they tend to live a long time.

Apple is an artificial monopoly. Their userbase is locked into expensive and proprietary hardware and software. That proprietary nature ensures no competition -- they control everything (hardware and software) about the market they serve. All they have to do is provide good quality and service; since they do, they are still around. Linux could probably maintain their own position in a niche market as well, that is, if someone else doesn't come along and offer a better open source alternative. As a Linux customer, who do I talk to to ensure the product I use is going to be maintained 5 years from now? It's a risk, maybe a small one, but still a risk.

Maybe, maybe not -- educate me. The biggest question mark in my opinion is wether ext (or any Linux FS) will be interoperable with Windows in the future. I'm sure there will always be a means to convert if you look hard enough (you brought up SAMBA -- that may be a possibility), but the question is one of useability.

I was talking about Linux drivers for NTFS (several existing drivers were mentioned in this thread already). If virtualization does in fact become the norm in Windows, and if Linux support is maintained, then yes, that will go a long way towards improving compatibility (on PC's). However, there is still the question of devices other than PC's...

Actually he was peculiarly unspecific about what he meant by "long term storage". Not presuming to assume any specific usage scenario, my recommendation also took into consideration that he might want to access his "long term storage" files using next gen portable devices -- devices that one can interface directly with a computer, or independently through a standard interface (like a HighDef video camera accessing files on a USB key or HDD).

The driver already exists and works well (ex. NTFS-3G), and by reasoning I already gave, future drivers will likely continue to.

Well, if you're talking about only using files on a your PC, and only about the lifetime of your current PC, and the Windows ext driver is stable, and particularly if Linux is your primary OS, then I would agree that the ext driver is the better solution. However, that scenario is what I would consider to be normal use, not "long term storage".