On March 13, a special report published on News.com site described the revival of an old Microsoft initiative. That News.com report said: “Microsoft is replacing the plumbing of its Windows operating system with technology borrowed from its SQL Server database software. Currently, documents, Web pages, e-mail files, spreadsheets and other information are stored in separate, mostly incompatible software. The new technology will unify storage in a single database built into Windows that’s more easily searchable, more reliable, and accessible across corporate networks and the Internet.”[…] I personally have written several letters to antitrust officials begging them to force Microsoft to publish the specifications of the Microsoft Office Application file formats. Such publication would really level the playing field, and allow users to decide on whichever office productivity applications they liked best. This, in turn, would give people much more flexibility in choosing operating systems. But just think — what if there were no file formats to publish? “Sorry judge, we would like to, but the data is not stored in files. It is stored in a database that is an indivisible part of the operating system.”
Read the editorial at NewsForge.
Our Take: While this filesystem will indeed introduce difficulties to alternative OSes regarding mounting and compatibility with “older” file formats, on the other hand, Microsoft should have the right to innovate. And as the article says, the file system is “in a single database built into Windows that’s more easily searchable, more reliable, and accessible across corporate networks and the Internet”. So, why the 94% of the desktop consumers should not be able to enjoy this easily searchable and more reliable filesystem db? Is it just because other operating systems (the rest 6% overall) will not have access into it? And at the end of the day, should this “other” 6% have access into it? Giving read (or write) filesystem access to a misson critical Windows system, it may not be a good idea after all. Surely this move will hurt the alternative OSes (but I just don’t believe that they started the SQL filesystem for this purpose, the db idea was around at Microsoft since 1993), but on the other hand, Microsoft is not bound and should not be bound by rival’s needs. Why should they?
Now, what Microsoft should do is publish the data structures of this new db filesystem, so other OSes will be able to modify e.g. PostgreSQL or mySQL to be able to read and write in this new filesystem. But holding Microsoft off from innovating in order the open source OSes to be able to play “catch up”, is just ridiculous and not fair. Not only unfair for Microsoft, but also for this 94% of the userbase, which translates into several millions.
it is a thread to open source because those gnu-ers won’t be able to plagiarize off MS and other companies.
This really means nothing for where open-source OSs excel – in the datacentre. A shiny new DB-driven FS will do nothing to dissuade me in my belief that windows has no place in serious MC applications. It can stay on the desktop, where most people prefer it. While I personally like NTFS and its VAX-like acls, this is certainly an intriguing bit of technology. However, ext3/vxfs and nfs/samba are the proven technologies that I’ll stick with.
Like anything, if this works out, it’ll be sure to be copied. ext4 with mysql backend, anyone?
have a db instead a filesystem isn’t exactly what I call a move for the speed…
And I don’t care so much about that stuff, still I’ll have my fun at a database crash…
There still will need to be file formats – people will want to send each other Word documents, etc. Given enough effort, it will still be possible for Linux users to exchange files with Windows users (or more frequently, to work on their own files under either Windows or Linux). It could, however, spell the end of the ability to mount Windows volumes in other OSes, or at least might require some really serious reverse engineering to develop the tools to do so.
Say, isn’t this “Cairo’s” OFS, from way, way back? For those too young, Cairo was the codename for NT 5, or what is now XP, a unified kernel for consumer, workstation & server OSes. The OFS, Object File System was a db file structure that would turn the volume into essentially a huge docfile. The idea was to achieve a sort of OpenDoc effect.
Don’t forget: at the same time MS was touting Cairo they were promoting “Blackbird”, which many thought had died, but is in fact Active X.
This doesn’t threaten OpenSource projects in anyway I can see. Besides, MS is under no obligation to keep going with FAT/NTFS so you can have an easier time with Linux.
I think there are alot of reliability and maintainability issues MS will have to put to rest before they can pull this off. Meanwhile, if the opensource community would jump on this idea, they could easily beat MS to the punch.
This sort of filesystem is the natural evolution of computers, I think the open source OS developers should embrace it (of course, GNUs HURD isn’t even production yet and still based on a FS).
Additionally, any new system that I buy can have its hard drive easily swapped out with a new one, and I can install any OS on it want. Why is this a problem for Open Source? It’s not like timid users are ripping off windows and installing linux now anyway…
David
i fail to see the advantage of that file system over say… ntfs
looks to me more like a way to force upgrades
I thought 2k was nt 5 (which was supposed to unify nt and 9x, but they weren’t quite ready so they slapped together me).
The BFS (BeOS File system) was a kind of database oriented FS. I’m not an expert in that field, but that was my impression. I see the new Microsoft FS as a sort of evolution from this one.
Personnaly I don’t see very much a treat to Open Sources, at the end is just another FS. I just hope it’ll not be too complicated to be reverse-engineered, so open source versions of this one will be fast developed for various OS, keeping easy the file sharing between two different OS …
“Currently, documents, Web pages, e-mail files, spreadsheets and other information are stored in separate, mostly incompatible software.”
Could have fooled me… I thought they were stored in my hard drive.
Applications interact with the filesystem to perform their I/O tasks. This isn’t changing.
What is coming is another OS-level service that allows you to work with files and the filesystem like a database. The filesystem is still there. File formats aren’t going away. NTFS isn’t going away.
Look at Index Services in Windows 2000 for an example. It’s a technology that apps can use for searching file contents. It does not replace the filesystem; it simply complements it.
Well, Microsoft will not break the legacy of backwards compatibility. It’s a benefit that keeps people wanting to use Microsoft’s OS. So I think it’s very safe to assume that legacy methods of accessing the filesystem will continue to work.
However, and this is only speculation, methods of accessing the new features of the filesystem could be kept secret. It’s habit of Microsoft to release a low-performance subset of their API. I don’t see any reason for this to change.
If they wanted to hurt Open Source software on Windows they could release another API (they’re a dime a dozen) under a non-disclosure agreement to certain companies. This is just another paranoid theory though I wouldn’t put it past them.
Microsoft is not bound and should not be bound by rival’s needs. Why should they?
So your choice is to help monopolies. This is a *political* point of view (refusing anti-trust protections), you want to help MS to remain a monopoly ?
If not, if as you say what matters is users and innovation, ask for public domain or BSD licenses for this fs. Not only data structures : other OS developers don’t have to work with MS abusive behavior. Data structures are insufficient. Look at what is done with Word or web browsers. Interface is public (.doc, .html), but they introduce proprietary elements other developers have to work with. MS APIs are always wrong, or not followed by their own softwares. APIs are only useful from honest and competent companies. They have shown who they are and what their weapons are. And they have proven they are dishonest, but you find easier to forget that ?
So either choose to help this monopoly and destroy any non-MS technology, or ask for *real* interoperability.
Reiserfs is on the same road actually, if you read their site you would know that. Only difference is that the Reiserfs developers know what they are talking about, unlike Micro$oft. Microsoft will probably convert the filesystem into a single registry-like database, making a single point of vulnerability just like the registry is today. Imagine simple file corruption resulting in the whole database filesystem going amok and losing all your data as a result.
>Only difference is that the Reiserfs developers know what they are talking about, unlike Micro$oft.
Riiiight…
What should we do dak completely strip ms of its rights to create patents (technically a violation of its corporate rights, which is a major lawsuit for any government official dumb enough to try it.)? No matter how much you don’t want ms to succeed/want a viable alternative you can’t take away their rights. To do so isn’t making a level playing field, its just changing who has the high ground. Besides if you do that what happens when the competition catches up to microsoft (I don’t think it will be linux because in order to catch them would require a single corporate face, and a product your trying to sell) then it has no reason to innovate because their the new king of the mountain, so 15 years from now your saying the same thing about the new monopoly.
>>Only difference is that the Reiserfs developers know what they are talking about, unlike Micro$oft.
>Riiiight…
Haha! that’s funny! Don’t you love how people think Microsoft is really stupid? I’m no fan of their practices, but they certainly are far from being technically incompetent.
I work on AS400’s all day every day, and the filesystem of an as400 is DB2 the database. and its very good.
but… the data we store on it is records..not files. so.. for every day use, collection of family photos, collections of music, collections of documents, great.. what about when I want to update 3 bytes in the middle of a ‘record’ (1 5mb executable)… hmm. issues. indexing! ouch.
still, i think a hybird 50-50 ntfs/db type file system is required…
well, its nice to see that microsoft is trying to incorporate something that the BeOS w/ its BFS had 10 years ago…. but i dont know about the ‘new technology’ label they’re putting on it.
-Kevin
I don’t care of Microsfot success or Linux as an alternative. What I was saying was for monopolies, and any competitor, includind proprietary software.
Now you’re talking about corporate rights, is it just because they exist there’s nothing to do, do say ? I hope you’re not one of those law fanatics. Companies have rights, these rights should be considered granted and should be removed from criminal companies. They can be a monopoly, (even more, 99.5%), I dont’t care, if interoperability exists. It means if they do not abuse. If there’s no viable interoperability (this means abuse), I don’t see why “rights” of a criminal company should be considered. It’s easy for them not to abuse.
And if it’s not a typo you talked about “patents”. Actually they shouldn’t exist, for any company, in a sane country. So this problem does not concern only MS.
In both case (government helping monopolies, and patenting ideas) I don’t agree these kinds of totalitarism.
“So your choice is to help monopolies. This is a *political* point of view (refusing anti-trust protections), you want to help MS to remain a monopoly ? ”
Actually, yes. Being in the IT field, if MS is in control, that means only one basic operating system I have to keep track of (Windows). Of course, there are differences between 9x/NT/2k/XP, but they are all the same in more ways then they are different.
I shudder when I think of the day that MS competitors dethrone MS by way of litigation, then 5 years later we have 5-6 different operating systems, each with 10-15% market share that we have to choose from, and more than IT personel have to keep track of. As if trying to keep up with Microsoft and its technologies wasn’t enough work – and that’s just one OS!!!
Not only that, but going to CompUSA and seeing a neat piece of software, and then finding out that it runs on each of the 5-6 operating systems on the market, except for the one I’m currently running. Ditto for the hardware.
So is MS being a monopoly such a bad thing? I personally think not.
The point of the article is that you will not be able to get files out of the file system (out of windows) any more and that is why open source will suffer.
I find that conclusion to be rediculous. You will always be able to get the file out of the file system so you can put it on a zip disk, cdrom, zip file for archiving, email, internet web site, etc. There is no reason to believe that you would not be able to get the file out.
The point of the db file system is too increase performance and maybe save some wasted hd space. I think the article does nothing to help anyone, makes no good points and should not have been published.
Or else the US wouldn’t have a law against it! I agree that we shouldn’t go ahead and kick MS in the nuts just to “even the playing field,” but I think they should take whatever punishment is warrented for the laws they’ve broken (and oh, yes, they’ve broken laws). As for the monopoly laws themselves, they’ve worked for more than a hundred years now, and I doubt they’ll fail this time.
It should also be pointed out that the little icons on your deskop only represent a “file” and that if you move the “file” or change the “file” that you are actually changeing the data stored in a file system. A database has the same outcome only it does it differently.
People understand the file “metaphore” and there is no way ms or anyone else is going to change the way that works (you will still be able to click and drag your “file” icons around the screen)
MS has the right to innovate. Why is the Linux community always playing catchup or bashing their innovations. Why don’t they innovate something revolutionary for the Linux OS instead of always copying Windows. Be Original come out with something first and let Microsoft and other OS’s play catchup. Linux community just bites other OS’s and duplicate it. Revolutionize the OS market with something wonderful, new and exciting put other OS’s to shame. THINK!!!
later
I only see lies in your post.
Linux and BSD systems already have lots of features Windows systems do not have. Actually they do not copy Windows, and most Windows/MS features are copied from other, older OSs or softwares.
Innovation is not a problem. And the Linux community does not bash MS for innovation. The only problem is interoperability.
Many operating systems doesn’t mean incompatible systems. Many softwares compile and run on many operating systems, it does not mean writing a new version from scratch for each system.
You’re in the IT field ? So what about applying monopoly to you ? If MS decides to do what you currently do, good bye. And what about applying this communism to every field ?
You’re right, totalitarism may be easier for profit fanatics, if it’s not applied to them. With a short term egoist point of view, it works, sure.
A database is in itself a file. The primary purpose of a database is to allow the storage and manipulation of data without having all said data in memory at any one time. Also, almost any filesystem can be thought of as a database already MSSQL, Oracle, MySQL, SysBase, InoBase… they are all general purpose databases to allow ease in creating database applications. They are not the definition of what a database is.
As I said before, any filesystem is a database already, and thus nothing new is being done with this “new” filesystem from the always “innovative” Microsoft. Almost everything Microsoft produces and sells, has already been thought up and researched by some other entity. The only original ideas that come out of Redmond are bad ones. One example is that the operating system should look and act like a web browser. The thing to remember is that the MS engineers faults, it’s the marketing department’s.
MS has the right to innovate<p>
For one, I’m getting a little sick of Microsoft, and all it’s clewbie hangers-on throwing this term around everywhere as if Microsoft was a fountain of brand new, never-been-done-anywhere-before ideas. From XENIX to DOS, to Windows (badly) aping the Mac interface, to NT’s VAX roots… well, I could go on and on. But mark, history has shown that if theres one thing Microsoft is good at, is taking technology, and usually mangling it. “Embrace and extend”. Remember these words. Remember Kerberos, as well?M$ has never done an original thing in their entire existance. This filesystem included, if it ACTUALLY gets off the ground this time. (a previous writer was correct – the OFS was destined for inclusion in a previous OS, but by MS’s own admission they simply could not make it work).
The main ‘innovation’ issue here is that it may be as fast. That’s the point that killed BeOS’s database FS (at least, that’s what I’ve read). Ideas are cheap, and if Microsoft pull it off (yet to be seen) then that’ll be impressive. ReiserFS is already out and has much of this. Microsoft’s .NET doesn’t allow one to access filesystem metadata, but I guess this’d fit more into the ADO.NET model. Microsoft are falling behind, at this point. I hope they catch up whenever they do release it.
Well, then you get the Postgres vs MySQL people arguing over the definition of a database. A filesystem is a catalogue of data but I think that the idea of a database filesystem can be defined as arbitrary data fields in the file system. This is something that current Microsoft file systems can’t provide (though Reiser does to some degree).
<p>Basically, the reasoning is that filesystems have been too lazy for too long. When a programmers writes an email program they may want to store 100 emails but they have to invent their own filesystem inside a file for speed purposes. Basically, the idea is to open up files to the filesystem again. That Reiser whitepaper has some excellent ideas on this.
<p>So, I think using a loose definition you could say that any filesystem is a database file. But I think it’s probably defined by arbitrary data fields for a file.
“Haha! that’s funny! Don’t you love how people think Microsoft is really stupid? I’m no fan of their practices, but they certainly are far from being technically incompetent.”
I don’t think Microsoft is stupid, quite the contrary. I think that people who rush out and buy the lastest MS operating systems and programs are though.
“MS has the right to innovate. Why is the Linux community always playing catchup or bashing their innovations. Why don’t they innovate something revolutionary for the Linux OS instead of always copying Windows. Be Original come out with something first and let Microsoft and other OS’s play catchup. Linux community just bites other OS’s and duplicate it. Revolutionize the OS market with something wonderful, new and exciting put other OS’s to shame. THINK!!!”
Well, since I can do a lot of things with my Linux machine that I can’t do with my Windows machine. And since I can do everything on my Linux machine that I can do with Windows, I would have to argue that Linux isn’t playing catchup and that Linux is, at least to some degree, original.
In my opinion, Linux has already put Windows to shame. That’s why I’m typing this on my Linux machine and my Windows machine only gets turned on when I’m being paid to do something on it.
I’m not sure why this is a threat to open source, but something like this is very overdue for mainstream computing. Files are so incredibly… 60s. The Newton had a database for it’s storage, and it was fantastic. We’re not talking about BFS, something way beyond it. Why most developers cling to them is beyond me- a simple object/struct-database can make things easier for everyone, for both end-users and developers.
I plan on using an OODB for my Squeak PDA Environment, and never considered creating new binary file formats, and why the Linux PDA folks insist on using such is beyond me. With a consistent db-as-fs system, it’s easier and more straight forward to create new apps which store data, and easier to sync. No need to create conduit plugins on the desktop and device side, the system just knows about the data contained in the db, and can sync therein.
Why are people all wyning here about MS innovation. Tell me one thing they invented?!? All there core stuff is coming from open source platform (bind, tcp/ip stack etc etc). And let’s talk about there innovating Activer Directory. Ye right a pure rip off of NDS. But giving the same features a different name makes it a innovation!
ReiserFS 4.0 will be out in some time. I bet MS is using coding from it …
Like every other IT company MS innovates something, they have their research and their UI design groups, etc.
I believe a lot of talented people work at Microsoft.
Their marketing department do what every other marketing department in the same situation does, their task is to maximize profit given the situation, and the situation is that MS has a monoploy on workstation OS and Office programs, hence they charge as much as they think they can get away with, practice dirty business, etc.
I think any large company would have done the same , think SUN, IBM, HP/Compaq, etc .
But it’s funny with all the pro MS people, cmon if you work at Microsoft I can understand you, but why support a monoploy company that charges you large amounts of money for ttheir OS and want to charge you much more, trying to introduce subscription services, etc which will make you pay rmicrosoft like you pay the rent.
Thinking is a process done by individuals with the help of their OWN brains.
“I personally think not” has to be processed by your own brains in the context
as you have put it in.
So, IF(not when) you ever start to THINK …. just let me know.
I also have to tell you that a lot of people never think. For now you are a lot of people. However there is always hope if you start to THINK.
>charges you large amounts of money for ttheir OS
we don’t buy their os, we “borrow” it from a warez ftp
the article expects m$ will allow access to this file-system
only with m$-products.
currently they can make interoperability hard by undocumented file-formats.
then they can block it by having to talk to the database,
means a running windows with it.
further a protocoll which is known.
will it be? otherwise m$-office can access the db and nobody else.
saving as files would then be very uncommon, no one would know how
(people don’t even know word can save html).
its even possible there are no files anymore, instead tere is kind of distributed db with parts on your and parts on my computer and these db’s talk directly.
propritary protocoll, 5 patents, no access from the outside then.
i don’t know if m$ will do it this way, but would be surprised if they don’t.
rezi
let’s all switch to pen, paper and a calculator
BFS was a journaled file system, but the contents (of the files) weren’t contained in a database as Microsoft seems to be implying. The simple fact remains that this project is a long way from completion and an SQL based “filesystem” would slow down the system greatly. It seems to me that MS is tring to have everything saved to one great database file, but down below would have to be a filesystem to handle that one file. It would be like combining files in a zip archive, it still needs a fs to reside on. Microsoft doesn’t know what it’s doing in this department and there is already a GPL’d journaled filesystem (atheos.cx). For some reason people want to make Linux viable on the desktop while AtheOS, a os that just needs more developers, goes unnoticed.
wouldn’t it be easier to mount the partition as a database and not have an extra fs layer to worry about??
Technically, yes NT5 is Win 2000. However, the original plan was to unify the DOS kernal & microkernal architectures in 1999, in NT 5. That was originally Cairo, as envisioned in 1993.
The Object File System is indeed the db file system M$ is now touting. For more info:
http://news.com.com/2009-1017-857509.html
Filesystems are databases — however, they don’t have all the auto indexing features. Databases are also probably cluster un-aware, and can probably make better use of disk storage. Not to mention, MS can kick start an entire new cottage inudstry for the enterprise market — High performance backend database filesystems for NT. It will allow them to simply put the MS badge over high performance vendor storage systems, my guess anyway, all living happily under the Windows roof.
Also, this will make having a Itanium or Sledgehammer a pre-requisite for the Windows with these features. That way, the kernel can just map in the entire partition on a local drive (or remote drive for that matter).
Its just better — but given some of the history points listed here — not sure that its so original.
David
uhhmmmm: I think you should check a dictionary, invention and innovation are two very different words, look them up before you use them next time, and maybe you won’t look like an ass.
Seraphim: And anyone cares because why? NTFS is a journaling filesystem too, big deal, that’s not what is being discussed in this article.
Whats next dak taking away criminals rights? And which criminals them all or just murderers and rapists? You are talking about a very slippery slope, and I for one hope we never go down it. you see dak in america everyone (corporations and all business entities count as well) has rights, and no one can take them away, if someone does they become the criminal. If you want to live in a country where rights can be taken away, go to a dictatorship (cuba is nice this time of year).
I fail to see any significant advantage for the average desktop user. All the things I can imagine that this piece of vendor lock-in can provide is already provided for by existing techniques. From a technical point of view this will only increase the complexity of the filesystem layer and introduce more possibilities of data corruption.
Data representation != data structure…
Actually these are not “rights”, but “privileges”, and privileges have to be taken away from anyone (not fundamental rights etc, of course, that was perfectly clear).
I don’t ask anything beyond what an anti-trust law should request. The problem is that anti-trust laws think with traditionnal industries, while IT competitors can’t live without interoperability. You’re talking about dictaroship ? Today MS dictates, and competitors have no right to exist withouT MS cryptoformats.
But of course, you’re trying to sound naive, you had understood I meant privileges, not fundamental rights. Continue your argumentation : next you’ll say anti-trust laws are criminal rights I suppose.
oops s/rights I suppose/I suppose/
Do you agree ms owns windows? If you do then do you concede its their right to change it as they see fit? Because what your saying is even though its their product its not their right to change it, without giving away the changes to the competition. btw your the one sounding naieve, you say ms dictates, it does have the major market share but they don’t dictate. If they did they could decide if anoyone else makes an os or office suite, they don’t, they just make their format proprietary. Its similar to copyrighting an invention, they don’t loose that right because they’re a monopoly. Even the government (through the antitrust laws) can’t request what you want them to, which is the loss of basic rights. It is the same as taking away a criminals rights, you are combining politics with business (btw if what you wanted could happen it would be communist not capitalism, and communism only works under a dictatorship, like the former ussr).
For a more detailed look at this subject, it is worth checking out “The Register” at the following locations:
Breaking the news August 2001: http://www.theregister.co.uk/content/archive/20879.html
Confirmation January 2002:
http://www.theregister.co.uk/content/archive/23852.html
Laughing at C|Net here:
http://www.theregister.co.uk/content/4/24534.html
Cheers
You know, if this were actaually an original idea, regardless of how BAD an idea it is, maybe you could call it “innovation” with a straight face. But its neither good nor origional. Just another blatant attempt to lock everybody into M$ crap for another five years. The most absurd part of it all is how tactics such as this hurt the industry. “Diversity? TRUE innovation? Nooo, we don’t want that!,” cry the paper MCSEs. Forget the fact that more divergence of technology means more opportunity for IT. Forget the fact that M$ has bought or stolen every GOOD idea it ever had. No, lets keep putting our trust in an abusive monopoly that was proven to be so in a court of law. That makes much more sense.
Many applications already use databases as filesystems: Documentum, FileNet and Clearcase are 3 examples that come to mind.
The first 2 actually are based on Oracle.
With Oracle your DB can be on a raw volume, no underlying FS structures there, just the DB itself… this is for speed so that you don’t have to go through the FS and the associated OS overheads. Works pretty well, actually.
Advanced databases can do nifty things like seamless clustering, replication, rollbacks/forwards, proper transaction handling (the next step up from journaling), easy recovery from crashes (if you take care of things properly), excellent indexing etc, to name a few.
Of course, there is a penalty for all this functionality and plain old FAT or any other simple, old FS that performed fantastically on 8-bit CPUs will probably always be faster for simple, normal operations than pretty much anything else but of course you won’t be able to do anything fancy like the things outlined above.
D
They own it, they can do whatever they want if they keep it.
But if they release it, they have to handle their monopoly and face it. They are a monopoly => They have to allow interoperability. Any proprietary format, when you’re a monopoly in a market that needs interoperability to have competitors, is a monopoly abuse. Only neoliberalism can say it’s not. They do dictate if they use proprietary formats. They are responsible for what they do, what they release. That’s the facts.
Is this similar to the BeOS filestsystem?
Many applications already use databases as filesystems: Documentum, FileNet and Clearcase are 3 examples that come to mind.
Here’s an easy one, the file format that Outlook stores it’s files in is a database. In fact, most of the Office file formats store a lot of extraneous information about the file and format within the file.
With Oracle your DB can be on a raw volume, no underlying FS structures there, just the DB itself… this is for speed so that you don’t have to go through the FS and the associated OS overheads. Works pretty well, actually.
MS SQL Server can do the same thing.
Advanced databases can do nifty things like seamless clustering, replication, rollbacks/forwards, proper transaction handling (the next step up from journaling), easy recovery from crashes (if you take care of things properly), excellent indexing etc, to name a few.
I think a lot of that, especially indexing (and especially in relation to metadata) is the reason that MS is trying to do this (and has been for 10 years). NTFS is starting to get a bit on the slow side as MS piles more metadata onto it, especially when dealing with directories containing a lot of Office or WMP-generated files (WMA or MP3, notably), and people would like functionality from Windows and Office that would probably easily justify rewriting the file system rather than adding anything else on top of the existing system. Too many home users use ‘Find’ instead of ‘Explorer’, and Find is just not good enough when you’re working with large volumes in FAT or NTFS, especially with people that aren’t quite sure what that file is called
It only makes sense to allow standard things like SQL or whatever to manipulate the database, as in all other databases.
At the very least there will be a set of APIs and all will be well.
It just means that people will have to write extra drivers in order to read the FS, that’s all. Actually, it might prove easier than reading NTFS…
It’s a good move. ReiserFS is going that way too, for your information.
BeFS was SUPPOSED to be like that but never made it completely (for performance reasons).
As computers get faster and faster, it becomes far easier to implement previously very expensive (in processor cycle terms) solutions. It’s all about functionality.
Look at MS Office. You might not like the fact that it’s a MS product but it IS highly functional, intuitive and complete.
BTW, I’m a professional Unix admin, not an MS zealot.
D