“To achieve the long-elusive goal of easily finding information hidden in computer files, Microsoft is returning to a decade-old idea. The company is building new file organization software that will begin to form the underpinnings of the next major version of its Windows operating system. The complex data software is meant to address a conundrum as old as the computer industry itself: how to quickly find and work with a piece of information, no matter what its format, from any location.” We have already reported on this here and here, but News.com has more information on the subject today.
New Windows Could Solve Age-Old Format Puzzle–at a Price
2002-03-13 Windows 29 Comments
Mainstream computer users could have all been doing this years ago if MS didn’t have a stranglehold on the desktop market. This sounds like a combination of BFS’s attributes system and grep.
All of the security holes will be labeled in 16 bit true color, with the font of your choice!
Read the side column of the news.com article:
“1992: Jim Allchin details a vision for a new version of Windows, code-named Cairo, which would include a revamped user interface and a new data store, called the Object File System (OFS), for storing document files, spreadsheets, multimedia files and other information.”
Microsoft was working on this for 3 years BEFORE Dominic Giampaolo (creator of BFS) even joined Be Inc. (in 1995).
It’s precisely that because Microsoft has to make the system somewhat backward compatible to DOS (and its millions of users), that mainstream computer users have not benefited from it 10 years ago.
Cairo is nothing else but Windows 95. It was Win95’s codename back in 1993-4. It seems that OFS was never used, but it was always under the “wait for now” state.
now we HAVE to upgrade to the latest MS crap….wel I guess I will be an exclusive Macer and Linux dude since all windows programs will no longer run on my win 98lite setup.
no, cairo was the code name for Win 2k / the culmination of the NT serise.
It’s just a rape of ReiserFS, NFS etc etc. And i would doubt it that the OS would be more stable. And if they want to use a dbase system they need a client, and what if the clients crash? I think OFS is an utopia. Hell what kind of hardware does that requires. And btw NTFS is good enuff, not the fastest, but stable.
And ofcourse it will be close-sourced which will make projects like Samba impossible. This whole damn plan has nothing to do with searching faster etc etc. It’s just a way to kick competitors out of the way.And it could even contain DRM.
…and for Windows 95, the code name was “Chicago”. I still have the BETA CD’s on my shelf.
AS/400 (iSeries) had this concept for years. It is basically a database on a disk. Everything had to be renamed: Mount point->Library,Directory->Table,File->Record.
it was a pain to move from Win32 and unix “file” to AS/400 record so they just created a filesystem emulator and a qsh shell.
BeOS & NT BOTH have already have free-form attributes and grep.
Don’t believe me, try:
To be honnest I’m not too sure what this OFS is? The stuff descused in the OSDev forum? Or just what they have with a db running the backend?
Back in the mainframe days, files tended to be more structured – fixed/variable record sizes were common and it was a hassle to move files around, backup, debug etc. After a while people figured out that making a file a linear sequence of bytes was much better to manage – let the apps do as they will with the files. By introducing structure again, I can see all these problems coming back.
My guess is that the windows registry and dlls have turned out to be an installation disaster and that pushing this kind of data out to the file system may be what’s required to fix it. The concept has its good and bad sides. I frankly don’t think MS have the technical foresight to pull it off correctly and we’ll end up with a horrendous forward/backward compatibility nightmare that will take years to recover from.
The idea of database style file systems is not new, and internally NTFS already has many pieces ready to roll. The real clincher is likely to be how such a file system appears to the applications. If past history is anything to go by, the API is likely to be a dog to work with – presumably .NET is what this is all about.
Here is something i think microsoft will not be able to do even with lot of money. We will see more file coruption than ever.
From what i did read of it, it make no sense. How can you have file represented in another way than a data stream on a winchester drive? It will have to be an abstaction layer to current FS model.
This is only a plan to hide the window FS to the competition. They will make it spagetti enough that by the time linux will read it they will already be at version 3. Now, just add the anti piracy harddrive stuff at hardware level and you see bill brain at work (in other word dedicated window HD, can’t change them).
.. is something like DTSearch, tied into the file system, so that any change of the FS triggers index updates. This does not break backwards compatibility, but makes searching gigabytes of different files (except for image searches) trivial.
I’ll be happy if Microsoft pulls it off. Linux will be quick to copy it if it works. Although “locate <bar>” and “grep -r <foo> *” are good enough for most stuff I do, especially when you’re not dealing with proprietary file formats 😉
hmm, sounds like old good Amiga datatypes system, which was in some form accepted by BeOS too IIRC 🙂
Just to be clear, it’s not just x=y metadata, it’s also namespaces and relationships.<blockquote>It’s precisely that because Microsoft has to make the system somewhat backward compatible to DOS (and its millions of users), that mainstream computer users have not benefited from it 10 years ago</blockquote>That doesn’t explain why metadata isn’t accessible via the .NET framework.
“We’re going to have to redo the Windows shell; we’re going to have to redo Office, and Outlook particularly, to take advantage” of the new data store, Microsoft CEO Steve Ballmer said in a recent interview with CNET News.com. “We’re working hard on it. It’s tough stuff.”
What’s that suppose to mean??? Is he trying to say that software built for earlier versions of Windows won’t work on Longhorn?
There has been no reason to buy new versions of most of Microsoft’s products since 1997. MS Office may be MS’s biggest revenue-earner at present, but they know that they have hit a brick wall. With Office XP, the only new feature that was worth advertising was that there was no more Clippy (that stupid animated paperclip). Even that was a lie: MS simply turned it off by default.
The solution: release a new OS which cripples the functionality of old apps. MS will advertise like crazy, hyping the ‘enhanced performance’ and ‘extra functionality’ of the new apps (which will be exactly the same, just not crippled like the older versions). But here’s the catch: you won’t be able to buy this on store shelves. This will be a subscription-only thing, part of MS’s .NYET strategy. People (particularly businesses) will be strong-armed into subscribing by MS’s licensing policies.
IIRC, Amiga and BeOS have had database filesystems, and that was five to ten years ago. Of course, they were too slow for the commodity hardware of the time, but things are different today. http://www.namesys.com/“>ReiserFS Reiser4″ rel=”nofollow”>http://www.namesys.com/v4/v4.html”>Reiser4 will have an extensible architecture, allowing all sorts of plug-ins. This will allow a database to be easily integrated into the FS. It is due on 30 September 2002, long before Longhorn is due.
I will believe it when I see it. Microsoft is notorious for putting out a crappy product the first two times. My prediction is that it will be released in late 2004 as an option. Dynamic partitions in NTFS will be the default…which will be most used except in the enterprise on select systems that will need to hold the indices for massive searching. The overhead will be too much for the average desktop or server for a couple more years.
Eugenia, you wrote:
> Cairo is nothing else but Windows 95. It was Win95’s
> codename back in 1993-4.
Sorry but that is incorrect. Win95’s codename was Chicago.
What’s that suppose to mean??? Is he trying to say that software built for earlier versions of Windows won’t work on Longhorn
No, he’s saying that they’ll have to redo it to take advantage of the new filesystem. Meaning they’re not going to ship the filesystem until Office has squeezed everything it can out of it. Remember that the most popular email clients for BeOS stored the email messages in the filesystem, whereas Outlook stores them in it’s own database format. So, to take advantage of a database-style filesystem, you drop the database that stores Outlook email and store the messages in the filesystem. Most of the Office formats are basically a filesystem in a file, and they can get rid of a lot of the overhead on the formats if they have this database style filesystem. The key, of course, is having Office ready to go when (or before) it ships, so that very few people have a chance to be on the market taking advantage of the oppurtunities the filesystem presents before they do.
If Office isn’t ready, you delay Windows, no big deal because Office is the money maker. Tell them you’ve found some security hole in Windows and that your new focus on security is so important that you’re holding Windows until it’s fixed. That should do quite well, and maybe release a few dozen patches for the beta program, maybe even patches that actually fix some things, to make it obvious that you are working on Windows.
I also forgot to mention, this will mean that the new version of Office will perform best on the new version of Windows, and that some features may rely on the filesystem so that they aren’t available on the old version of Windows. Therefore, if someone really sees an advantage in the features of the new Office, they’ll have to upgrade both Office and Windows to take advantage of it. I wouldn’t be surprised if they also updated a lot of the ‘integrated’ apps in Windows to take advantage of it, such as WMP, and updated the Windows Media formats as well. I’m sure a lot of the updates, in addition to removing overhead, will be in how metadata is stored, because currently the Office formats store that information within the file, and the structure (as well as storage method) may have to be changed to work better with this filesystem (similar to the fact that WinXP treats ID3 tags on MP3s as metadata).
little off topic well very, but Yama clippy is not turned off by default. He is there from the start and is his same old evil self.
In one move MS will break all OSS and commercial software that tries to read and write to MS file systems and applications. You can bet they will encrypt this file system as well so that if anyone
tries to reverse engineer it they will sue them into oblivian using the DCMA.
Looks to me that MS is going to make a gamble, If you want to use MS software you are going to have to be 100% MS or none at all. All the way or not at all.
No more Linux/BSD, Solaris, AIX, etc networks. MS on the desktop and MS on the server or none at all. Want to read that word file you just got? Well you can forget about using Star/Open Office, Productive etc. You will have to have a NEW copy of MS office, which will only run on the NEW version of MS windows, Which will require a NEW PC.
Their new filesystem will the SQL Server Yukon with filesystem semantics, and compatibility drivers a la M: drive like in Exchange. Nothing will break except for disk checking and defrag tools. Also it’s no rape of ReiserFS & co, coz I’m not aware them being built on a database, or do you want to tell me ReiserFS is a cloaked mySQL? Also I don’t see how a new filesystem automatically generates more file corruption! My running SQL Server instance didn’t corrupt any data till now and prolly never will.
> little off topic well very, but Yama clippy is not turned
> off by default. He is there from the start and is his same
> old evil self.
Okay, so then the MS advertising for Office was not simple deception as I had alleged, but much worse: blatent lying. In that case, what point was there to buy any MS Office product after ’97?
office 2000 had a better “feel” to it than ’97 but if I had already had ’97 I prolly wouldn’t have bought 2000, so the difference is pretty negligible.
I haven’t purchased new MS products (or stolen them, for that matter) since my free upgrade to Win98FE on the day it was released. I see no reason to buy or use anything else from there on. If an alternative OS does not come along that I am comfortable with, I will stick with my 233MMX Win98FE/BeOS R5 machine and be content or start painting on cave walls and recording with pro audio hardware that doesn’t need a computer. I wont spend money or effort upgrading to new Microsoft market-capture ploys. MS stuff isn’t even worth stealing any more.
Oh yeah… that’s just me. The brainwashed masses will eat it up and the rest will follow out of “compliance.” Oops. Will the world never learn…
Mea culpa.. I wasnt happy with Win 98 stability and 2000 didnt satisfy my software requirements so I downloaded and installed XP Pro…. I know a lot of Pcs that are perfectly happy with Win98 and run for weeks without a glitch.. mine wouldnt, no matter how much I tweaked it, so I went for the unsupported free upgrade route!!!!!
Ultimately, people are always going to do this until they find they have software that suits their needs… and until the price of a PC Windows upgrade gets more reasonable ( OSX from OS9 is a much bigger performance leap than 98/ME to XP and half the price ) you’ll always have piracy.
I now have a PC which has both compatibility, speed and reliability.. I have no compelling reason to upgrade my OS again. Which doubtless represents a problem for MS – what killer app will they pull out of the hat to sell their next OS, now that they finally got it right with XP in terms of usability and reliability.????
A new File system, even if it is a decent Journalled one with metadata, wont be enough for me or most other people.