For a flexible task-based interface (one where a user would be able to have choice), the user would have to be bombarded with way too many questions. Going through these choices is much more time consuming than an interface that allows the user to be the boss.
The world isn't ready for a task-based interface unless MS manages to pull off a Star Trek-like voice-activated task-based system. Then I'd be really impressed. Anything less would be the death of MS.
You are not looking at the whole picture here, are you?
Longhorn is scheduled for 2006. We actually have no idea what the interface is going to be like, except that the technical underpinnings will be 3D accelerated like in OS X. If the past is a good judge, they will make a passable interface that people will be content to use because they don't know of anything better.
As far as I know, MS is doing a very big rewrite of the Windows codebase. It could mean a real enhancement in security and stability. If they do it well enough, and compromise on backwards compatibility (they have already shown this in XP SP2) Longhorn can be about as reliable as any other OS.
Then there is WinFS, which IMHO is a very big thing, has enormous potential. MS has a lot of database experience and a good track record at building them, so I think WinFS will work in practice as well. (Actually, the biggest reason it will work because it needs to work 100% if they are going to get an OS running on top of it.) This is something Apple does not currently have an answer for, technologically. It's questionable how useful MS will make this for the mainstream user and how soon, but if they manage to do it, everyone else is in big trouble until they can respond with similar technology. There are some seeds of that sort of development going on around Linux - ReiserFS, Gnome Storage for example. But they are pretty far from what WinFS promises to be.
OS X version (insert number here) is scheduled for 2006. We actually have no idea what the interface is going to be like.
It is simply ridiculous to compare an OS two years out to one we have now that will surely be updated by then.
Direct comparison is ridiculous. What I said was, it is important that there is major improvement in OS X when/if WinFS really takes off and gets app support, otherwise it's hard to compete.
No, WinFS need not work, because Windows does *not* depend on it.
WinFS is NOT a file system.
WinFS does NOT mean Windows File System.
You're probably correct in that Windows does not necessarily have to depend on WinFS. But if they don't trust it enough to do that, I think it means WinFS will be pretty much relegated to non-existence.
WinFS is a file system. If you mean the fact it's built on top of NTFS, do you also think that ext3 is not a filesystem because it's built on top of ext2?
WinFS does mean "Windows File System" (whoa, big surprise there):
Looks like they changed it; originally it meant "Windows Future Storage".
And no, I mean by that that it depends on NTFS to work. It won't work with FAT32 or any other file system, and it won't work on its own. It relies on certain NTFS features but its not its own real file system because it's optional.
And no, I mean by that that it depends on NTFS to work. It won't work with FAT32 or any other file system, and it won't work on its own. It relies on certain NTFS features but its not its own real file system because it's optional.
If you can access a WinFS disk with NTFS drivers and doing it doesn't mess up WinFS, then I don't consider WinFS as a filesystem. Otherwise I do.
The issue isn't so much whether the technology is there as it is whether Microsoft or anyone else comes up with ways to use it that improve the user experience.
Apple's filesystems have had vaguely database-like characteristics from the get-go: The desktop database, metadata in files, etc. Currently, there's a sort of piecemeal approach where iTunes has its database and iPhoto has its database and so on. This allows Apple to figure out what concrete functionality a database layer should have, and it accustoms people to managing data through what amount to database views.
There is also the question of whether having a complete layer in place makes sense. Obviously, quality of implementation is an issue here, but for example why not offer a database as a standard set of Foundation classes that respond to a standard interface, and let applications use that? This formalizes Apple's current approach of application-centric metadata, so that it's easier to implement and easier to script (meaning that one application can ask another what it thinks is so special about file X). This approach just occurred to me, so I haven't hashed out the advantages and disadvantages completely. But it makes sense on a certain level to have information important to a given application be associated with that application; and it seems cautious to keep such databases small and independent for reliability and security reasons (we don't want the Windows Registry!); and there's already a robust system in place for querying other applications. It's not document-centric, but frankly I'm not sure how well that paradigm scales anyway. It's great for handhelds and appliance-like computers, but part of me likes the UNIXy idea that a file is just a file and it's the applications that give its contents significance.
On the other hand, if you do want to associate metadata with files, OS X is all set up for that. There's already a virtual filesystem in place, and bundles are just perfect for stashing metadata - people just have to get used to the idea that documents as well as applications can be funny-looking folders.
Good post and good points Amorph. Indeed, iTunes is probably the most widespread consumer app that has substantial database "feel" without being called a database.
There is a clear need for a systemwide database. One question I'd like to be able to ask the computer is "what did I do on 24.12.2003" and when I ask that, I want to see e-mails from that day, I want to see the playlist that was playing, photos taken, code written, et cetera. Until metadata support is ubiquituous, fast and standardized to a high degree, there is no chance of getting this feature. It would also help a lot to have most files as text (markup languages) so it's possible to index the contents as well.
Good post and good points Amorph. Indeed, iTunes is probably the most widespread consumer app that has substantial database "feel" without being called a database.
There is a clear need for a systemwide database. One question I'd like to be able to ask the computer is "what did I do on 24.12.2003" and when I ask that, I want to see e-mails from that day, I want to see the playlist that was playing, photos taken, code written, et cetera. Until metadata support is ubiquituous, fast and standardized to a high degree, there is no chance of getting this feature. It would also help a lot to have most files as text (markup languages) so it's possible to index the contents as well.
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
Why impractical? HD won't stop growing in size. One day we'll have storage devices that could store all of today's Internet content just to show you that technology will never stop advancing.
Logging this info will bring lots of flexibility (and complexities too) to the end-user.
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
I don't think so. Put together, my browser history for today, the text files I've written or modified, my e-mails and two different playlists probably take under 10kB, certainly under 100kB.
If we go to the other extreme, even if this feature somehow ate 100MB of hard disk space every day, a modern 250GB HD would still last seven years as a history disk. Long enough in my book.
The thing that is missing is the programs to support this. Hardware is ready and has been for some time.
Comments
The world isn't ready for a task-based interface unless MS manages to pull off a Star Trek-like voice-activated task-based system. Then I'd be really impressed. Anything less would be the death of MS.
"Our OS of 2 years in the future will beat your OS of today!"
That's why the word vaporware was invented for this company. Makes comparisons pointless. Thurrot is about as reliable (read: on crack) as MOSR anyway
Longhorn is scheduled for 2006. We actually have no idea what the interface is going to be like, except that the technical underpinnings will be 3D accelerated like in OS X. If the past is a good judge, they will make a passable interface that people will be content to use because they don't know of anything better.
As far as I know, MS is doing a very big rewrite of the Windows codebase. It could mean a real enhancement in security and stability. If they do it well enough, and compromise on backwards compatibility (they have already shown this in XP SP2) Longhorn can be about as reliable as any other OS.
Then there is WinFS, which IMHO is a very big thing, has enormous potential. MS has a lot of database experience and a good track record at building them, so I think WinFS will work in practice as well. (Actually, the biggest reason it will work because it needs to work 100% if they are going to get an OS running on top of it.) This is something Apple does not currently have an answer for, technologically. It's questionable how useful MS will make this for the mainstream user and how soon, but if they manage to do it, everyone else is in big trouble until they can respond with similar technology. There are some seeds of that sort of development going on around Linux - ReiserFS, Gnome Storage for example. But they are pretty far from what WinFS promises to be.
WinFS is NOT a file system.
WinFS does NOT mean Windows File System.
OS X version (insert number here) is scheduled for 2006. We actually have no idea what the interface is going to be like.
It is simply ridiculous to compare an OS two years out to one we have now that will surely be updated by then.
Oh look, bevelled buttons, embossed into a grey metallic looking window... Now where have I seen that before?
Originally posted by progmac
what is a task-based interface?
http://www.winsupersite.com/images/reviews/wme9_01.jpg
Originally posted by SilentEchoes
You know, Gon, That works both ways.
OS X version (insert number here) is scheduled for 2006. We actually have no idea what the interface is going to be like.
It is simply ridiculous to compare an OS two years out to one we have now that will surely be updated by then.
Direct comparison is ridiculous. What I said was, it is important that there is major improvement in OS X when/if WinFS really takes off and gets app support, otherwise it's hard to compete.
I think you misread my post.
Originally posted by Chucker
http://www.winsupersite.com/images/reviews/wme9_01.jpg
I really, truly, hate wizards.
Originally posted by Chucker
No, WinFS need not work, because Windows does *not* depend on it.
WinFS is NOT a file system.
WinFS does NOT mean Windows File System.
You're probably correct in that Windows does not necessarily have to depend on WinFS. But if they don't trust it enough to do that, I think it means WinFS will be pretty much relegated to non-existence.
WinFS is a file system. If you mean the fact it's built on top of NTFS, do you also think that ext3 is not a filesystem because it's built on top of ext2?
WinFS does mean "Windows File System" (whoa, big surprise there):
http://msdn.microsoft.com/library/de...ic_storage.asp
And no, I mean by that that it depends on NTFS to work. It won't work with FAT32 or any other file system, and it won't work on its own. It relies on certain NTFS features but its not its own real file system because it's optional.
Originally posted by Gon
I think you misread my post.
I think you are right
Originally posted by Chucker
And no, I mean by that that it depends on NTFS to work. It won't work with FAT32 or any other file system, and it won't work on its own. It relies on certain NTFS features but its not its own real file system because it's optional.
If you can access a WinFS disk with NTFS drivers and doing it doesn't mess up WinFS, then I don't consider WinFS as a filesystem. Otherwise I do.
Apple's filesystems have had vaguely database-like characteristics from the get-go: The desktop database, metadata in files, etc. Currently, there's a sort of piecemeal approach where iTunes has its database and iPhoto has its database and so on. This allows Apple to figure out what concrete functionality a database layer should have, and it accustoms people to managing data through what amount to database views.
There is also the question of whether having a complete layer in place makes sense. Obviously, quality of implementation is an issue here, but for example why not offer a database as a standard set of Foundation classes that respond to a standard interface, and let applications use that? This formalizes Apple's current approach of application-centric metadata, so that it's easier to implement and easier to script (meaning that one application can ask another what it thinks is so special about file X). This approach just occurred to me, so I haven't hashed out the advantages and disadvantages completely. But it makes sense on a certain level to have information important to a given application be associated with that application; and it seems cautious to keep such databases small and independent for reliability and security reasons (we don't want the Windows Registry!); and there's already a robust system in place for querying other applications. It's not document-centric, but frankly I'm not sure how well that paradigm scales anyway. It's great for handhelds and appliance-like computers, but part of me likes the UNIXy idea that a file is just a file and it's the applications that give its contents significance.
On the other hand, if you do want to associate metadata with files, OS X is all set up for that. There's already a virtual filesystem in place, and bundles are just perfect for stashing metadata - people just have to get used to the idea that documents as well as applications can be funny-looking folders.
There is a clear need for a systemwide database. One question I'd like to be able to ask the computer is "what did I do on 24.12.2003" and when I ask that, I want to see e-mails from that day, I want to see the playlist that was playing, photos taken, code written, et cetera. Until metadata support is ubiquituous, fast and standardized to a high degree, there is no chance of getting this feature. It would also help a lot to have most files as text (markup languages) so it's possible to index the contents as well.
Originally posted by Gon
Good post and good points Amorph. Indeed, iTunes is probably the most widespread consumer app that has substantial database "feel" without being called a database.
There is a clear need for a systemwide database. One question I'd like to be able to ask the computer is "what did I do on 24.12.2003" and when I ask that, I want to see e-mails from that day, I want to see the playlist that was playing, photos taken, code written, et cetera. Until metadata support is ubiquituous, fast and standardized to a high degree, there is no chance of getting this feature. It would also help a lot to have most files as text (markup languages) so it's possible to index the contents as well.
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
Originally posted by Whisper
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
Why impractical? HD won't stop growing in size. One day we'll have storage devices that could store all of today's Internet content just to show you that technology will never stop advancing.
Logging this info will bring lots of flexibility (and complexities too) to the end-user.
Originally posted by Whisper
That would indeed be cool, but I think that logging that much information on a personal computer is a bit impractical. You'd almost have to create and store a daily backup of your drives to achieve what you're talking about.
I don't think so. Put together, my browser history for today, the text files I've written or modified, my e-mails and two different playlists probably take under 10kB, certainly under 100kB.
If we go to the other extreme, even if this feature somehow ate 100MB of hard disk space every day, a modern 250GB HD would still last seven years as a history disk. Long enough in my book.
The thing that is missing is the programs to support this. Hardware is ready and has been for some time.