Mac OS: Next big step?
With the old OS´es (Dos/Win95 and Classic) everybody knew that the foundation was rotten and there was something better out here that "just" needed to be tuned to todays needs for user friendliness. With regards to Mac OS NeXT had already proven that it was possible, long time before it was brought into Apple.
But what about the next step WRT the OS? Steve called X the OS for the next 10-15 years IIRC and we are already a third into that. Any magic bullets out there just waiting to be picked up? Any new revolutionary concepts?
But what about the next step WRT the OS? Steve called X the OS for the next 10-15 years IIRC and we are already a third into that. Any magic bullets out there just waiting to be picked up? Any new revolutionary concepts?
Comments
Originally posted by Anders
....
But what about the next step WRT the OS? Steve called X the OS for the next 10-15 years IIRC and we are already a third into that. Any magic bullets out there just waiting to be picked up? Any new revolutionary concepts?
Call it what you want, but it is still ten years before the end of MacOS X in the most plausible accelerated timeframe. This works out to about 200 years in the life of a human being. You can rest assured that Apple, Microsoft, and others are researching new computer OS concepts. Slow evolution will continue until someone discovers new paradigms for interacting with our computers. Like most significant discoveries, new OS discoveries will not be made until we need them.
Originally posted by Anders
With the old OS´es (Dos/Win95 and Classic) everybody knew that the foundation was rotten and there was something better out here that "just" needed to be tuned to todays needs for user friendliness.
Uhm well, yes. OTOH, MacOS X builds on foundations that are nearly 40 years old (UNIX), so in a way, it was a step back in time.
Difference is: the "new" foundations are rock-solid since they have trickled down from main frame computing to PC's. The concepts have been refined over the years to be very flexible, extendible and to some extend fault-tolerant (old MacOS tended to crash when one app crashed, OS X - and WinXP - rarely do so since there are barriers between the OS and applications).
All really new concepts have failed, no matter if we are talking about Atari or Be.
Therefore, I don't really believe we are going to see another generation of OS foundations, most likely, the current structure in OS X, Linux and Win32 will continue to evolve instead of being replaced by something completely different.
I expect some major steps by DRM (isolating the OS from the user - evil, but it will come) and virtualization.
Think about how Spotlight enhances and modifies the desktop metaphor. Files are immediately at your finger tips because you no longer have to wade through folders and sub folders to find things. In the short term I think that desktop search will have a major impact on the development of OSs and in the long term, who knows what will emerge.
FTFF
Regardless of whether you move beyond the desktop metaphor a lot of people are still going to be using it so a nice and happy finder would be good and all would be right with the world.
After that a new file system to replace HFS+/HFSX
Linux has Reiser4 which is flat out awesome for example, also Spotlight could be integrated deep into a new file system to achieve more BeOS style. Not that this would be noticed by the end user except for bringing some snappy.
The next big that the user would notice?
No idea, I certainly haven't been impressed with 3D stuff (Sun's Looking Glass), nor Microsoft's liberal use of pointless transparency, and Spotlight needs a lot of work speed wise.
Originally posted by Electric Monk
The next big step?
FTFF
Regardless of whether you move beyond the desktop metaphor a lot of people are still going to be using it so a nice and happy finder would be good and all would be right with the world.
I agree. The Finder is lame and I think its the one of the weakest points of OS X. FIX IT APPLE! Use some of that innovation to give us something better.
This gives (assentially what Java was designed to do) you a lot of new options that are a bit tricky to implement today. For example, you could run front end of software on your computer and the back end on another computer over the network. Some applications do allow you to do that today, but in the "new world" any application will be able to do that.
If another hardware is better suited to your tasks, migrating to a new computer (say from Windows to Mac) will be trivial and you'll never have to get readjusted to the new "feel" of the OS.
Microsoft and Apple will not loose anything here since they will still be providing the underlying OS and making money off the software.
Some food for thought.
Originally posted by baranovich
I'm always wondering what will be in store for the future of OSs. The big question is what will replace the desktop metaphor???
Imho, nothing. The desktop is here to stay.
We had been discussing this among friends some 10 years back and we all believed that with the incredible power PPC was offering (at 120Mhz), the desktop metaphor could be replaced with something better... turns out it is a) good enough and b) something people are used to.
Originally posted by Smircle
Imho, nothing. The desktop is here to stay.
We had been discussing this among friends some 10 years back and we all believed that with the incredible power PPC was offering (at 120Mhz), the desktop metaphor could be replaced with something better... turns out it is a) good enough and b) something people are used to.
Apple is slowly transitioning to a metadata-rich environment (the process must be gradual or users will not adopt it)...the desktop as we know it with it's spatiality and it's hierarchical organization is on it's way out.
Instead of organizing things by folders in physical space, we're moving to organizing things by attributes.
The reason why some people hate Spotlight right now is that Apple isn't let people harness all of it's power yet. The OS and file-system are still heavily dependent on hierarchical organization.
The Spotlight implementation is still, at best, half-baked. The current Tiger Open/Save dialogs still only take into account that you're organizing things hierarchically when they could be much more by allowing you to tag metadata to the file. "This file is related to 'work'" "This file is 'schoolwork' and is part of 'show and tell' presentation".
Instead of going through a static folder layout, you'll be able to query for anything related to work...or anything related to schoolwork or anything related to the show and tell for the past year.
Hierarchical organization was fine in the past when the hierarchical trees were small due to the small number of files that needed to be organized. But the number of files on one's HD is growing fast nowadays. It's becoming much easier to lose files on your HD especially for people that are disorganized and don't want to spend time to organize their files. Hierarchical organization is becoming inadequate. It should still exist, but it should also make way for a database/metadata system as the default way of finding files.
The next big step is improving Spotlight...metadata tagging...the file system...and the query interface (the current Finder interface is a bit of a turn off for newbs.)
Originally posted by baranovich
I agree. The Finder is lame and I think its the one of the weakest points of OS X. FIX IT APPLE! Use some of that innovation to give us something better.
Amen!!!!!!!!!!!!!
What if the entire OS was built using a 3D metaphor instead of trying to add 3D manipulation to the current 2D desktop (like Sun's Looking Glass). The OS could be like a 1st person shooter 3D game (think something like UT2004) and interacting with the 3D environment would be analogous to manipulating 2D windows.
Files could be organized in various ways. They could be grouped into different kinds of buildings or different locations within the 3D world. Instead of remembering which folder contains your files, you could remember a spacial location within the 3D world. (A sophisticated search might be in order to prevent any files from being lost.) The files themselves could be represented as "statues" where a different type of statue represents a different type of file. To access a file you'd simply "walk" up to the statue and press a button to interact with it.
I know this idea is out there, but I figure that in order to replace the desktop, the replacement is going to have to be very radical. I also realize that you can't completely remove all 2D elements of the OS and that applications would still run in a 2D mode.
My imagination has been working overtime on this one. Thoughts?
Originally posted by baranovich
Files could be organized in various ways. They could be grouped into different kinds of buildings or different locations within the 3D world. [...] The files themselves could be represented as "statues"
If you think about it again: do you really, honestly feel this is going to improve your productivity or do you like it for the coolness factor?
Some guy has used the Doom-engine as a wrapper around Linux process control - you could see processes walking around and kill their avatars (thereby terminating the process). I liked it for its absurdity - but I have yet to see anyone who found this useful in his daily work.
Of course, there are ways to have a 3D interface that isn't based on spatiality (3D and spatiality aren't mutual exclusives) but the example given in this thread is an example of spatiality.
One of the largest attributing factors to this is versions. People tend to want a back up of the file before continuing work in case the file gets corrupted. A classical example lots of people have on their computers is a combination of several of the following:
- resume.doc
- resume1.doc
- resume2.doc
- resume3.doc
- resume.bak
- resume.bak1
- resume.bak2
- \\old\
esume.doc
- \\old2\
esume.doc
- \\bak\
esume.doc
- etc.
- etc.
I'd arugue that using metadata for versioning and somehow giving easy to use version control to end users would be a big win for the number and organization of files that we have.
Just look at an iPhoto library. It uses xml for a control database with lots and lots of files in a folder hierarchy to simulate versioning. iPhoto would take much less space if the filesystem stored the different saves of a file as binary diffs (like SVN does) between versions. And if the filesystem/finder could let a user grab any version of a file she chose, nobody would have to worry about needing special tools to extract older versions.
Another win would be having spotlight aware of the versioning system as well. You could search for a key word and spotlight could tell you that a year ago, the word was in file X. You could look at that old version or open up the current one...
Originally posted by rrabu
Regarding "people have lots more files on their HDs making organization hard":
One of the largest attributing factors to this is versions. People tend to want a back up of the file before continuing work in case the file gets corrupted. A classical example lots of people have on their computers is a combination of several of the following:
- resume.doc
- resume1.doc
- resume2.doc
- resume3.doc
- resume.bak
- resume.bak1
- resume.bak2
- \\old\
esume.doc
- \\old2\
esume.doc
- \\bak\
esume.doc
- etc.
- etc.
I'd arugue that using metadata for versioning and somehow giving easy to use version control to end users would be a big win for the number and organization of files that we have.
Just look at an iPhoto library. It uses xml for a control database with lots and lots of files in a folder hierarchy to simulate versioning. iPhoto would take much less space if the filesystem stored the different saves of a file as binary diffs (like SVN does) between versions. And if the filesystem/finder could let a user grab any version of a file she chose, nobody would have to worry about needing special tools to extract older versions.
Another win would be having spotlight aware of the versioning system as well. You could search for a key word and spotlight could tell you that a year ago, the word was in file X. You could look at that old version or open up the current one...
Certainly...with the capacity of HD on the rise, I don't see why the OS couldn't automatically keep several versions of a file and tagging it with metadata so that you can go back to a previous version at anytime. Especially for text files which take up a miniscule amount of space.
Originally posted by rrabu
I'd arugue that using metadata for versioning and somehow giving easy to use version control to end users would be a big win for the number and organization of files that we have.
That would be a godsend. Developers use version management systems since like 15 years or what. Having it for any kind of data and work transparently in the background would be a real step forward.
BTW: If you want a reason why I don't believe in 3D-GUIs, check out Microsoft Bob, one of the worst desasters in MS' history. It wasn't only ugly, but also completely confusing and always in the way. AFAIK it was the first and at the same time the last 3D-GUI ever...
No applications. Or these are invisible. Files are put on various places on the desktop, in specialized drawers... A document would be on only compatible format. The document would be visually seen as a piece of paper, or a photo, a piece of film, depending of its content (it could display a reduced preview of its content). You could take it and bring it back to front, zooming in it, zooming it out. There would be no optimal 100 % fixed view, all views would be good looking. You could write things on it with text processors pallet tools (think Word/Pages/etc), or draw things with drawing tools (think Illustrator or Photoshop).
The metaphor with reality would be a lot better.
Originally posted by Cosmos 1999
Bring back the old OpenDoc concept (dropped from Copland) and push it to the max.
No applications. Or these are invisible. Files are put on various places on the desktop, in specialized drawers... A document would be on only compatible format. The document would be visually seen as a piece of paper, or a photo, a piece of film, depending of its content (it could display a reduced preview of its content). You could take it and bring it back to front, zooming in it, zooming it out. There would be no optimal 100 % fixed view, all views would be good looking. You could write things on it with text processors pallet tools (think Word/Pages/etc), or draw things with drawing tools (think Illustrator or Photoshop).
The metaphor with reality would be a lot better.
This concept doesn't really work in the real world because companies want to sell apps, not invisible extensions to the OS. It would be beautiful if the concept did work though but it's like telling everyone to burn all their money because you don't need it anymore in a world where everyone is offering their services for free.
Originally posted by kim kap sol
This concept doesn't really work in the real world because companies want to sell apps, not invisible extensions to the OS. It would be beautiful if the concept did work though but it's like telling everyone to burn all their money because you don't need it anymore in a world where everyone is offering their services for free.
I know, I was talking about a perfect ideal world... :sigh:
Originally posted by Anders
But what about the next step WRT the OS? Steve called X the OS for the next 10-15 years IIRC and we are already a third into that. Any magic bullets out there just waiting to be picked up? Any new revolutionary concepts?
OS X uses the Mach Kernel, etc, and in at least that respect it is probably the most advanced OS in the world. It is very well suited to meet the demands of todays consumers and today's programmers. Windows is not, nor will Longhorn do anything -- as far as I can tell -- to fix that.
The next big step in OS's will occur with the next big step in computer usage or computer design. OS X was meant to address Internet, multimedia, and handling large amount of data on a CPU-centric system. Even Cell essentially uses a centralized "star topology." As far as the usage hold-all "using large amounts of data," I don't think that's going to change in the forseeable future, but the amount of CPU-based devices in a given environment will increase. It's possible that the next big OS development will be along the line of what FSMLabs has done with RTLinux. This would be useful in the sense that application units can run independent instances of the CORE, communicate will each other across variable interconnect, and all stay reasonably in sync without the higher-level code even realizing how the operations are being distributed. So essentially this is distributed computing and cohesion across various electronic appliances that may or may not be "part of the computer."
If this is how it happens (which I think it will), there's no place in the world for a proprietary OS such as Windows.