Maybe to you Classic was useless, but tell that to a business with tons and tons of files that could only be opened by a Classic app when Leopard came out. Same for Rosetta. Same for PPC. Just because you consider something to be useless doesn't mean it is...
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Is Apple running out of ideas?
- Thread starter Retina MacBook
- Start date
- Sort by reaction score
Maybe to you Classic was useless, but tell that to a business with tons and tons of files that could only be opened by a Classic app when Leopard came out. Same for Rosetta. Same for PPC. Just because you consider something to be useless doesn't mean it is...
It's completely fine for me, even to wait 4 years for the next, major release as long as it's shockingly amazing. Or just as amazing as Tiger - Leopard. The UI changes were Amazing!
No, technology is imroving in pretty high rates. Apple will always have something new to release. For example, have you ever seen a computer with a 2560x1440p resolution like iMac's? Even 3D TV's still use 1080p. Imagine that the new Retina MacBook Pro has a resolution of 2880x1800p. Can you imagine that. There's no other company that offers such an elegance. And no; I am not a fanboy.The freshness of a Mac feels great. In a few months we'll all see the revolutionary iPhone 5 (or whatever it's going to be called). Even Mountain Lion is quite amazing. It takes full advantage of the iCloud service, so that all of your Apple devices are always synced. Don't forget the built-in social network services like Twitter and Facebook.
The resolution of the display, in case you didn't notice, isn't about software, it's about hardware. You can put Mac OS X on a TV and that wouldn't make it 2660x1440. But, If you put a different OS on an iMac, provided that the drivers are installed, it will run 2660x1440.
Who camps outside Apple for software updates when they can be readily downloaded from the app store?
I don't see what's wrong with refinements. If the OS itself is pretty solid as is, there is little point in trying to reinvent the wheel. So it is just introducing additional features to fine-tune those existing features.
I am pretty interested in some of the touted features. Icloud seems like it could make for a worthy dropbox alternative (because I can sync documents directly). Imessage, well, fun to contact someone's iphone/ipad while working. All these small little stuff that improve my workflow without any real learning curve or changing the way I use OSX. What's not to like?
That wasn't his point. He opened his statement with "technology is always improving" which is counter to your question "is Apple running out of ideas?"
If 'running out of ideas' means a yearly update with fewer features to enable this then i'm Ok with it.
If it also means not coming out with something as ill-thought out as Windows 8 is then i'm OK with that as well.
I was always previously a windows sympathiser I guess you could call it, ie until Win 7 I always saw a place for it. I've tried windows 8 and *HATE* it - so I think that Apple are getting something right with their release scheme that none of their releases are as terrible as that.
If it also means not coming out with something as ill-thought out as Windows 8 is then i'm OK with that as well.
I was always previously a windows sympathiser I guess you could call it, ie until Win 7 I always saw a place for it. I've tried windows 8 and *HATE* it - so I think that Apple are getting something right with their release scheme that none of their releases are as terrible as that.
I don't think iOS has anything to do with the amount or lack of new features OSX is getting... (especially since iOS isnt getting much "big" features this time either, except maybe the new map thingy. )
And i'm pretty sure those are 2 seperate departements that dont have to share alot of ressources.
"Big" things just take longer to develop than 1 year. I'm sure Apple is already working on some major changes for OS XI (or whatever it will be called). Hopefully ready in 1-2 years.
And i'm pretty sure those are 2 seperate departements that dont have to share alot of ressources.
"Big" things just take longer to develop than 1 year. I'm sure Apple is already working on some major changes for OS XI (or whatever it will be called). Hopefully ready in 1-2 years.
At least Mountain Lion doesn't have all of the goofy colorful tiles only a kindergartener would enjoy.
However I agree, Leopard was a big step forward for the OSX. After Leopard, Apple's upgrading the OSX is marginal at best. However Apple just wants to integrate their IOS and the OSX now. I think that's why we're not seeing a really big innovative upgrade. Apple just wants everything to work well together on all of their products.
However I agree, Leopard was a big step forward for the OSX. After Leopard, Apple's upgrading the OSX is marginal at best. However Apple just wants to integrate their IOS and the OSX now. I think that's why we're not seeing a really big innovative upgrade. Apple just wants everything to work well together on all of their products.
comments on XNU development
First off of the 3 major kernels XNU (OSX), NT kernel and Linux; XNU is far and away the least sophisticated. Apple has always been behind on kernel related issues from 1997 till now, though in 1997 they were closer to Linux. There is nothing new about that.
Here is an example where Apple tried, but tried and failed. They were aiming for ZFS which while not exactly innovative would have been a major step up. There is some 3rd party work that looks promising: Greenbytes / Zevo
I'm not even sure what you mean here. Things are much better in the drivers department than they were a decade ago, when I was using wrapped Linux drivers.
I'd say the capacity to use an integrated and a discrete graphic subsystem and switch seamlessly based on power is a rather major advance. And that was only possible because of the Open CL in 10.6 WebKit2 layout engine (10.7) was a key advance for graphics in practice.
What does that even mean? Apple has had full support for v6 for years. IPv6 subnets are supposed to be global. There is no reason for a small private network to use v6. The reason to use v6 on a private network is that every device has an externally addressable IP.
First off of the 3 major kernels XNU (OSX), NT kernel and Linux; XNU is far and away the least sophisticated. Apple has always been behind on kernel related issues from 1997 till now, though in 1997 they were closer to Linux. There is nothing new about that.
Here is an example where Apple tried, but tried and failed. They were aiming for ZFS which while not exactly innovative would have been a major step up. There is some 3rd party work that looks promising: Greenbytes / Zevo
I'm not even sure what you mean here. Things are much better in the drivers department than they were a decade ago, when I was using wrapped Linux drivers.
I'd say the capacity to use an integrated and a discrete graphic subsystem and switch seamlessly based on power is a rather major advance. And that was only possible because of the Open CL in 10.6 WebKit2 layout engine (10.7) was a key advance for graphics in practice.
What does that even mean? Apple has had full support for v6 for years. IPv6 subnets are supposed to be global. There is no reason for a small private network to use v6. The reason to use v6 on a private network is that every device has an externally addressable IP.
XNU kernel
Sure. Take an example of filesystem support. On XNU you have a very basic 1970s style filesystem underneath the covers. That's why you start having allocation problems as you get bigger than 1TB, which people are hitting. You also have inconsistency issues about case sensitive throughout the OS. It lacks advanced features for large drives so that block level corruption in HFS+ creates file corruption, which with larger files is a problem.
And mind you, you don't even get a high speed filesystem, (like a modern XFS) would provide.
Linux has support for well over 100 filesystems and there is 3rd party support on Windows for every filesystem I've ever run across. So not only is the filesystem choices meager on XNU, the default kinda sucks.
____
In terms of protocols there is a move towards move advanced protocols like SCTP to replace TCP and UDP (why SCTP). Again most kernels support this already.
As an aside, the Windows Networking stack with multiple points for runtime configuration and virtualization is just amazing. The result are Windows applications that are able to reconfigure the networking stack in minor ways without having performance implications or serious compatibility problems.
____
I personally think the /proc filesystem and the ability to view in a reasonable and accessible way kernel status information is incredibly powerful. There is no easy way to figure out what's going wrong with XNU kernel if it is having problems: (description of proc). Admittedly Windows has nothing like this either.
____
Now most of the advanced features of the Linux and NT kernel mainly apply to servers and not desktops. But here XNU started life as a server kernel. So it doesn't have embedded features (especially useful for phone) like you would find in a real time kernel. For what's its worth, I think given that Apple doesn't make server products (in any meaningful sense) anymore, they should switch to a real-time kernel. Desktop users would much rather trade off slightly lower total work for a system that is 100% of the time responsive to new input. The system, while being slightly slower, would feel so much faster. I don't know whether RIM will survive long enough to show off the advantages of a real time kernel but I hope so. I think the idea of having a server kernel ported to the desktop and then reported to phones is a terrible idea. I understand why Apple did it, and the advantages of being able to port over the entire Cocoa framework made it possible to get off the ground quickly.
Sure. Take an example of filesystem support. On XNU you have a very basic 1970s style filesystem underneath the covers. That's why you start having allocation problems as you get bigger than 1TB, which people are hitting. You also have inconsistency issues about case sensitive throughout the OS. It lacks advanced features for large drives so that block level corruption in HFS+ creates file corruption, which with larger files is a problem.
And mind you, you don't even get a high speed filesystem, (like a modern XFS) would provide.
Linux has support for well over 100 filesystems and there is 3rd party support on Windows for every filesystem I've ever run across. So not only is the filesystem choices meager on XNU, the default kinda sucks.
____
In terms of protocols there is a move towards move advanced protocols like SCTP to replace TCP and UDP (why SCTP). Again most kernels support this already.
As an aside, the Windows Networking stack with multiple points for runtime configuration and virtualization is just amazing. The result are Windows applications that are able to reconfigure the networking stack in minor ways without having performance implications or serious compatibility problems.
____
I personally think the /proc filesystem and the ability to view in a reasonable and accessible way kernel status information is incredibly powerful. There is no easy way to figure out what's going wrong with XNU kernel if it is having problems: (description of proc). Admittedly Windows has nothing like this either.
____
Now most of the advanced features of the Linux and NT kernel mainly apply to servers and not desktops. But here XNU started life as a server kernel. So it doesn't have embedded features (especially useful for phone) like you would find in a real time kernel. For what's its worth, I think given that Apple doesn't make server products (in any meaningful sense) anymore, they should switch to a real-time kernel. Desktop users would much rather trade off slightly lower total work for a system that is 100% of the time responsive to new input. The system, while being slightly slower, would feel so much faster. I don't know whether RIM will survive long enough to show off the advantages of a real time kernel but I hope so. I think the idea of having a server kernel ported to the desktop and then reported to phones is a terrible idea. I understand why Apple did it, and the advantages of being able to port over the entire Cocoa framework made it possible to get off the ground quickly.
I would argue that the EXT line of filesystems is in a very similar situation and given that it is the default on almost all of the widely used distributions, this should apply to Linux.
Are you referring to volumes or files?
Granted, I spent only a minute or two thinking about this but a noticeable omission is ZFS. I'm pretty sure I could come up with more if needed.
I also think it is not correct in a discussion of Kernel features to include those provided by third parties. These filesystems drivers you mention are not part of the NT Kernel.
I think it would be extremely naive to think that SCTP will ever replace TCP/UDP in even the distant future ono anything except private networks perhaps. It's been RFC for more than a decade and what's the rate of implementation in the real world? Seems kind of academic at this point.
I completely agree with this.
Correct me if i'm wrong but NT is not realtime either?
OS X has had support for preemption and schedulers for a long time. This article is from 2001:
http://static.usenix.org/events/bsdcon/full_papers/gerbarg/gerbarg_html/
Linux did not achieve a lot of these kinds of features until PREEMPT was added years after in the 2.4.x line of kernels.
PREEMPT_RT is still not in the mainline kernel source and needs to be patched in:
https://rt.wiki.kernel.org/index.php/RT_PREEMPT_HOWTO
Whilst I agree that Apple no longer makes server products (at least none serious enough to warrant major kernel level support), I disagree that XNU is a pure server kernel. It is not even close.
Again, I disagree with your assessment of the current state of support for desktop features in XNU (see above article), but I think you're missing a bigger point. Prior to PREEMPT and the desktop style schedulers by Con Kolivas (-ck kernel patches), Linux on the desktop felt slow. OS X in it's current state does not and for this reason I feel that you are calling for a fix to a problem which does not exist.
To me this sounds almost identical to the transition Linux made from its server roots, to the desktop and then to Android.
Linux distribution are mostly on ext4 at this point. That's the 5th major filesystem revision in 20 years, with almost constant improvements ever week. I don't think the situation is similar with Linux at all. Just a quick look at ext4 features I think makes the differences rather clear.
Well first off the only really good ZFS is running under Solaris on Sun/Oracle hardware. Also ZFS is a Unix filesystem so Windows would also be treating this as non native. In which case, iSCSI offers ZFS for Windows. There is a google project right now to implement ZFS for Windows.
Microsoft has always had a 3rd party development model. I don't want to bias the conversation about advantages by getting into the the "who pays" model.
In 2010 I was shocked how little worked had happened for IPv6. Today there are major IPv6 implementation efforts with every carrier. By 2015 I expect to see a complete flip.
No it isn't. But NT has to support a line of server products.
I agree 100% that Linux is worse in this regard than XNU, which has been pointing towards desktops since the NeXT days. NT was built from the ground up with nice adjustments for the desktop. Linux as you mention is just getting those.
That being said though. I get, especially prior to my recent hardware upgrade, unacceptable stalls all the time with XNU. This isn't something that doesn't exist its part of my daily working experience.
I agree. iOS and Android are in the same boat here.