I would've thought otherwise. I see your point regarding throughput, but part of the problem with a traditional hard drive is the need to move the head to access those tracks. A hard drive can read sequential tracks nicely, but what about reading 50 such tracks concurrently? The heads are going to have to move, negating the sequential read advantage.
The answer to the concurrency questions is "more spindles".
Actually, using something like iRAM would be even better for streaming. Though I wouldn't entrust it with a live database, serving video from it would be ideal.
SSDs will stream quickly, no doubt about it (although not _that_ quickly - most seem to top out in the 50-60M/sec range). But their maximum size is very small - especially if you want to stay within the realms of nominal affordability.
Standard drives and RAID will deliver a much better bang/buck for media streaming than SSDs will, unless you're streaming relatively tiny media files (in which case you're probably still better off spending the money on more RAM for your server(s)).
He's going to be shocked to find out that Vista is a DRM hell hole. Maybe Steve never told him.
Vista's "DRM" isn't any different to XP's (or even 2000's) "DRM". It won't magically DRM-encumber to your currently unencumbered media (unless you tell it to of course).
If you don't want to be restricted by DRM, don't buy DRMed content. Follow that rule, and Vista is no more "DRMed" than any other platform.
This would make an ideal drive for streaming media servers [...]
No it wouldn't. Even ignoring the pitifully small size of SSDs, the advantage to them is in latency, not throughput. You want them for things with lots of small, random seeks (DBs *are* a good example), not big, long, continuous reads like media streaming.
The Real World is not about the good guys fighting the evil villains in third world countries, no matter what your american tv tells you.
I'm not American and about the only American TV I watch are "Numb3rs" and "Battlestar Galactica".
In reality it's composed of a multitude of small evils, one of them is clearly microsoft.
Sure. So is a drunken bloke groping a drunken woman at a party.
However, out in The Real World, there are so many other, more important "corporate evils" to be concerned about - in particular ones where actual people suffer actual harm due to events outside of their control - that the SlashGeek call to arms over "Micro$oft is teh 3v1l !" doesn't even pass the laugh test.
Well, if you've never read a book about logic, maybe it does.
Maybe if you've never read a book about logic, "I've never seen Windows win in a head-to-head competition" carries more weight than "I've seen numbers that indicate it might handle CPU sharing as well or better" as well.
Simple fact is that Windows has been tuned and developed for much bigger systems than OS X runs on, for much longer. Its current implementation has distinct technical superiorities to OS X's (eg: more fine grained kernel locking, more mature Async I/O) that result in it scaling better.
It applies to all cursor controls I've used mice, tablets, trackballs, pads, joysticks, even. If you haven't run into this, I can't imagine you've ever done graphics work.
I haven't, nor did I ever say that I did. I *have*, however, interacted with people who do - and none of them use the mouse for "drawing lines".
Yeah everyone is making it up because we're all part of a conspiracy... that's it.
Oh, grow up.
I didn't say anything about a "conspiracy", I said Mac users have been complaining about Windows's mouse tracking since day dot and *in my opinion* it's because of the different acceleration algorithms the two OSes use. I can relate, since the mouse acceleration in OS X (and Classic) used to annoy the hell out of me as being too slow. Now I just adjust my technique to take advantage of the different acceleration curves, but it's still noticably different for a good hour of use after moving from Windows to OS X.
A lot of Windows users get shoved in front of a mac, use it for a while, and then realize the problem they learned to live with was not an inherent flaw in all devices but a Windows bug that has never been fixed. Or it's all some form of mass hallucination, you pick the more likely one.
I use Macs, Windows, Linux and Solaris quite frequently. I have never noticed _any_ "mouse tracking bug" in Windows on the scale you are talking about, compared to other platforms. I *have* noticed that they all have different pointer tracking characteristics, and I *have* noticed people from one platform complain about that - but that's just a matter of adjusting to something different.
I didn't know you were referring specifically to UI responsiveness, but actually I remember seeing something (at arstechnica I think) a while back that did just that on a macbook running Windows then OS X.
Well, I'm not sure how else you could interpret the term "sluggish" in the context of multitasking, but that's not really important.
Ars are good for that sort of thing, but Siracusa tends to only compare to earlier versions of MacOS [Classic], not Windows. Some results from Google, like this one, do not do a lot for supporting any assertion that OS X is faster (although I believe there's recently been some patches that specifically address WoW performance).
Hahahaha! That is why it is better for power users. You see for larger numbers of windows, picking the application then the windows results in many fewer key presses and UI interactions. Think 10 applications each with 10 windows. To get from app 1 window 1, to app 5 window 5, with windows you have 50 key presses.
Or one, if it was the last window you used...
To get there in OS X you have 10 key presses (5 to get to the right app, then 5 to get to the right window). This only applies to larger numbers of windows and has a lower learnability than the Windows version. For power users who do it a lot, changing the key you press is not really any slower but pushing it an addition 40 times is.
Changing the key is important because it's not just changing the key, it's changing the mental model of what you're doing from "finding the right application" to "finding the right window". In Windows, the cognitive load is lower because you only have to be concerned about "finding the right window" - the applicatio
My father once told me: "you cannot be neutral between good and evil".
Absolute judgements are rarely accurate.
Not to mention, there's a vast, vast difference between "good and evil" and "something and Microsoft". On the scale of corporate bad behaviour, wouldn't even make it past the cutoff for the bottom quarter.
Anyone who calls Microsoft "evil" without their tongue firmly buried in cheek, *seriously* needs to get of their mother's basement, take a reality check and get some perspective on The Real World.
Maybe they're thinking that Apple didn't have a bad idea with OS X... but where Apple went with Mach and a BSD userland, Microsoft could take a Linux kernel and then wrap an interface and a Windows API compatibility layer around it.
To what benefit ? There's little - if any - technical advantages the Linux kernel has over the Windows kernel. Indeed, many of the things that are "new" in the Linux kernel, the Windows kernel has had for a very long time, if not forever.
Of all the things Microsoft need to "fix" in Windows, the kernel is a long, long, long way down the list.
What about in ten years, when said Slashdot Geek has long graduated from university and gets promoted to a position where they can start to make purchasing decisions? (Apologies to those with 4-digit UIDs, who may already be old enough for this to have happened.)
By that time, most of them will no longer *be* the Slashdot Geek.
Ok, then you agree that there is something unique and interesting with the finder. Great.
I agree that they're useful. They're not "unique and interesting" - firstly because they're not unique and secondly because they've been around for nearly a decade now, putting them well and truly out of the "interesting" phase.
But nobody said anything about it being new. No argument there.
The OP was specifically referring to the timeframe while Microsoft has been working on Vista. Which - at worst - starts around late 2001. Specifically, he said Apple "introduced a new method of file browsing (love it or hate it, Finder is unique and interesting)".
(There were other errors in his post as well - eg. about OS X being "totally new", but the Finder one was the most conspicuous.)
Now as to replacing it, that is what spotlight will hope to be doing. Finder is certainly better than spotlight at this stage of the game. Definitely better than window explorer. At least I can drop to a real prompt and make any hierarchy I want with hard and soft links.
Comparing Spotlight to Explorer (or Finder, for that matter) is like comparing Word to Excel - nonsensical. While there's a slight overlap in functionality, they're meant for different things.
Services are no good if there is user interaction. I'm talking about applications. Tray icons are a hack and confuse most users to no end. They don't behave as users expect and don't have any way to provide the normal menus. Small windows are the common hack used to get around this problem. I've seen plenty of programs that have small, useless windows that take up space but are otherwise useless. So yeah, there are ways to work around the problem, but they are just nasty UI hacks.
I must admit I'm struggling to see what type of application you're talking about that needs user interaction, but can't justify a small window to put it in. It's not like Windows or OS X aren't already rife with these sort of "applets", nor that they violate existing UI guidelines.
I wish that were the case, but it just isn't so.
Well, I wish OS X was as fast and responsive to use as Windows, so I could use it full time without becoming frustrated - but it just isn't so.
I've seen numbers that indicate it might handle CPU sharing as well or better, but in the real world with disk and memory and other bottlenecks I've never seen Windows win in a head-to-head competition.
What "competitions" are you talking about ? Windows runs on much, much bigger systems than any version of OS X does and has been doing it for a lot longer. That alone, suggests it is more mature at utilising hardware resources and scaling to them. Heck, it's only in the latest versions that OS X has _started_ to implement fine-grained locking in the kernel (and is not even close to completing it). From a scalability perspective, OS X would barely be on par with Windows 2000.
One of the reasons macs are so popular among graphic professionals is if you're drawing a line on an OS X system it is unlikely that a background application will ever take enough resources to make the mouse input be ignored by the system or not displayed with the cursor. This is a common occurrence on Windows systems. Maybe this will be fixed with Vista as I have not tested it extensively yet, but for WinXP, Win2K and earlier this is certainly the case.
Can't say I've ever experienced it, but then again I rarely draw lines with the mouse. Nor do most graphics professionals I've ever met, either, they use graphics tablets.
Mac users have been complaining about Windows's mouse tracking for a couple of decades now. Personally, I think it's just the different methods in mouse acceleration rather than any real problem.
Your first statement is not supported by the benchmarks I've seen.
What cross-platform benchmark measures UI responsiveness ?
Both OS X and Windows are faster at different tasks with different applications. Now that both run on the same hardware, this is even more apparent. As for switching applications, it is better right now on OS X than it is on WinXP and historically, it has been a draw depending upon your workflow and experience.
I find Windows far more efficient for my usage pattern, which is lots of separate windows being switched between relatively frequently.
For people using the mouse as their switching mechanism, OS X is a clear win. For people using the keyboard, OS X was behind up until a couple of years ago when they extended the keyboarding and now, it is maybe a bit behind for novices and a bit ahead for power users.
Er, no. Task switching in OS X via the keyboard is demonstrably slower because it requires multiple key combinations to do so. You're either using Cmd+` to move between two windows in the same application (same as Windows) or you have to use Cmd+Tab to move to the right application, then potentially Cmd+` to move to the right window. In Windows, it's a single key combination.
OS X's UI is application-centric. It's designed with the assumption that users will spend relatively large amounts of time using a single application, possibly interacting with multiple documents within that application, and
I also doubt that Oracle would be stupid enough to limit support for their DB to Unbreakable linux. Their DB is their bread and butter, if they drop support for any OS that they curently support (say RHEL), there will be customer attrition. Why would they want to do loose Database customers for the sake of a product (Unbreakable Linux) that they will not make as much money on?
For Oracle customers, the only point of interest is Oracle. It doesn't matter if it's running on Linux, Solaris, Windows or a room full of monkeys banging rocks together - as long as the millions of dollars they have sunk into their environment, plus the millions of dollars in generates in revenue, that relies on Oracle to continue running, does so.
There are very, very few Oracle customers who are in any position to move away from Oracle. If Oracle says their DB will only be Supported on "Oracle Linux", then "Oracle Linux" is what's going on the DB servers.
Ideally, every software should be implemented in a way to make such changes, if not trivial, at least possible, regarding time and cost constraints.
Nice in theory, but it inherently either prohibits any company from implementing software that does $TASK better than anyone else, because they are constrained by having to retain compatibility with all the software that doesn't do $TASK very well, if at all, or it ignores the possibility of any $TASK that isn't designed and specified by a committee.
Or, to put it another way, for the result you want, all "innovation" - especially from individuals/small groups - has to cease, *OR* every developer has to work in collaboration with every other developer and sacrifice their ability to create a competitive advantage.
In other words, it ain't going to happen. Ever. As you note, the real world brutally intrudes on academic theory at the point where real money gets involved.
On Windows, this type of application would confuse the crap out of people.
In Windows, this type of application tends to be either run in a service, minimise to a tray icon or simply display a small window (and is extremely uncommon, in any event).
Further, I often want to close all the documents in use by an application, but not quit the application. The reason for this is also simple, I don't want to wait for it to load again multitasking us good enough that having it open does not slow down other applications. I think this has also been an issue related to Windows since on Windows users have trained themselves to quit applications not in use, since the multitasking does not handle programs sitting idle in the background very well.
IME Windows multitasks much better than OS X does. Firstly because OS X is so sluggish to begin with - and adding more running apps just makes it worse - that interacting with multiple tasks is discouraged, and secondly because the task-switching UI - particularly historically - isn't as good. Expose certainly added a dramatic improvement on previous versions (and cleverly hides some of the sluggishness with flashy graphics), but it becomes difficult to use with numbers of windows in the double digits, especially if they are similar (eg: terminals) or screen resolution is relatively low (a problem with most Macs).
Interestingly (with regards to your comment) while Microsoft was doing UI research for Vista, they found that Windows users tend to "multitask" - that is, run more stuff at once - more than Mac users did. They also tended to switch between tasks more frequently. This was outlined in the Vista GUI blog that was linked from Slashdot a month or two ago. It also mirrors my experience, both as a user and observer and, in my opinion and experience, it's because Windows simply handles multitasking better and has done so for much longer. I certainly know that every Mac I've ever used has choked up under the sort of multitasking load and usage pattern even my old dual P3 could handle (running Windows or Linux). My 1Ghz iBook can't even handle as much stuff simultaneously as a Compaq 500Mhz P3 notebook, despite having ~30% more RAM and easily twice the raw processing power (although, obviously, individual processing-intesive tasks are much faster).
It pisses me off when I'm on a Windows box and I close a file only to have reopen the program that just quit because I want to open another file using that same program.
If you're immediately reopening an application it should be essentially instantaneous, since all the relevant data will still be cached - unless your machine is memory starved.
This is a common task that Windows application developers hack their way around by creating managing Windows that just sit around as placeholders with nothing in them, or useless features in them that only duplicate functionality of the menus.
What applications do this ? I can't think of any off the top of my head, so I highly doubt it's a "common task that Windows application developers hack their way around".
There is one difference and that is the dock. If you close a window the user will notice the program is still running pretty quickly, since the dock icon remains and is marked as running. Most novice mac users use the dock to start and stop applications. It makes a lot more logical sense, in my opinion, than tying the action to having open windows.
The Dock's indication of what is and isn't a running program is, IMHO, far too subtle. Certainly many novice users I know (including my parents) have trouble with it. Of course, IMHO the Dock as a whole is a UI train wreck, and a significant step backwards from the equivalent UI in MacOS Classic, and especially Windows. Certainly looks cool, though.
Every screenshot I've seen of Aero looks remarkably...Aqua-ish.
Can you think of a way to implement various things like transparency and fancy effects today that *wouldn't* look "Aqua-ish" ? Follow-on competitors are always going to "look like" the first product to market they're competing with to some degree.
Really, if you're comparing just the two of them, they're not especially alike. The way they use transparency, fading and other effects (like "sheets" rolling in and out of windows, or "Flip3D") is all quite different.
I've been using Macs for a very long time, and Vista daily for a couple of months now, and the two have vastly more differences than they do similarities - and most of those similarities are the kinds of things that can appear even with two completely independent implementations - just like MacOS and Windows have had, well, pretty much forever.
Not in the details, but I can't help thinking that someone at Microsoft took a look at Aqua, and decided that it was probably time to overhaul Windows' interface as well; not to mention doing the same sort of graphics-card offloading that Apple did with Quartz Extreme.
The Windows GUI has been getting the same kind of "overhaul" it gets in Vista with pretty much every release since Windows 95 (which was the last _major_ change). It's reasonably safe to say even without the pressure of OS X, it would have gotten a similar update to the one it actually did - if only because 99% of people (even on Slashdot) don't look any further than that to see what's "new and improved".
Moving the display system onto the video card is just a natural and obvious progression. Microsoft had been talking about doing it since before Apple did, although OS X obviously beat them to the market with an actual implementation. Saying they were "copying Apple" in this regard is specious, at best - about as accurate as saying Apple were "copying Microsoft" by implementing an OS with pre-emptive multitasking and protected memory.
Introduced a new method of file browsing (love it or hate it, Finder is unique and interesting)
There is nothing "unique and interesting" about Finder. In two of its three modes, it's just a (less functional in some ways) copy of the MacOS Classic Finder. In the third, it's a copy of the NeXT file browser.
(And they all suck, IMHO, but that's not reall the question here.)
Another of Apple's little fits of genius was to make the laptop automagically write the RAM/states to disk - essentialy switch to a hibernated state - when battery power drops to a dangerous level. Waking then takes a few extra seconds, but is still much faster than a reboot - and of course the user state is precisely as you left it. So closing the lid on an iBook effectively lasts forever, because Apple was smart enough to make the distinction between Sleep and Hibernate transparent to the user.
Which iBook ? My 1Ghz iBook doesn't do this - if the battery goes flat when it is sleeping, it does not resume, it boots to a fresh desktop.
Ah! So X is only half of a GUI. Gnome and/or KDE is the other half. That makes sense... kinda. Although it does raise the question "If X has been around for such a long time providing the bottom half of a GUI, why hasn't there been an equally seasoned top half?"
Becuase the only really useful feature unix nerds see in X is the higher resolution allowing them to fit more xterms on the screen...
You simply cannot spend 2 years cranking out a dot level upgrade to an OS and expect thundering success. Because in the final analysis there isn't much new to Vista
Vista is a solid 5.x ->6.x revision. The changes - particularly in the lower levels - are significant, on par with the updates Apple did to NeXT to get to OS X. That most people do not bother looking past the relatively minor GUI update does not change this.
I'm not a MS fan, or really anti-MS - (read my other posts) but Microsoft are stuck here, they cant simply adopt a radical new operating system (even with virtualisation) and throw out all legacy support. They have gone as far as they can with Vista (which really isn't a re-write, its a mash up and an attempt to address some of their major problems).
Microsoft's problem is not their codebase or technology, it's their development process.
With legacy support they will continue to have the same security problems that have plagued them since the inception of windows.
Any platform with the marketshare and userbase Windows does will have the same problems Windows does.
The reasoning is simple, if Microsoft adopted a *nix like kernel and re-wrote everything then they would have an OS with little or no software support, little or no Hardware support, and they would find themselves competing directly with Linux / Solaris / BSD etc...without the benefit of already having a huge installed user base with a clear upgrade path. There would be one additional exception, their offering would not be free (unless they would simultaneously go down the software as a service route and make their OS free.. which throws up even more issues see My other post)
They'd also be taking a technological step backwards or, at best, sideways.
The Windows NT kernel is about the *last* thing Microsoft needs to be changing.
Sorry, but if Apple, with way less resources, was able to completely change their OS, change platform and still provide a migration path for developers, why microsoft, with vastly superior resources and cash, would not?
Because firstly, and most importantly, Apple has fewer scruples about breaking/obseleting old hardware and software.
Secondly, because Apple's user base is an order of magnitude smaller.
Slashdot Mirror
I would've thought otherwise. I see your point regarding throughput, but part of the problem with a traditional hard drive is the need to move the head to access those tracks. A hard drive can read sequential tracks nicely, but what about reading 50 such tracks concurrently? The heads are going to have to move, negating the sequential read advantage.
The answer to the concurrency questions is "more spindles".
Actually, using something like iRAM would be even better for streaming. Though I wouldn't entrust it with a live database, serving video from it would be ideal.
SSDs will stream quickly, no doubt about it (although not _that_ quickly - most seem to top out in the 50-60M/sec range). But their maximum size is very small - especially if you want to stay within the realms of nominal affordability.
Standard drives and RAID will deliver a much better bang/buck for media streaming than SSDs will, unless you're streaming relatively tiny media files (in which case you're probably still better off spending the money on more RAM for your server(s)).
He's going to be shocked to find out that Vista is a DRM hell hole. Maybe Steve never told him.
Vista's "DRM" isn't any different to XP's (or even 2000's) "DRM". It won't magically DRM-encumber to your currently unencumbered media (unless you tell it to of course).
If you don't want to be restricted by DRM, don't buy DRMed content. Follow that rule, and Vista is no more "DRMed" than any other platform.
Of course, if Windows behaved properly, it wouldn't even write to the page file until AFTER the RAM was full (or damn near full).
Why do you believe this is "proper" behaviour ?
This would make an ideal drive for streaming media servers [...]
No it wouldn't. Even ignoring the pitifully small size of SSDs, the advantage to them is in latency, not throughput. You want them for things with lots of small, random seeks (DBs *are* a good example), not big, long, continuous reads like media streaming.
The Real World is not about the good guys fighting the evil villains in third world countries, no matter what your american tv tells you.
I'm not American and about the only American TV I watch are "Numb3rs" and "Battlestar Galactica".
In reality it's composed of a multitude of small evils, one of them is clearly microsoft.
Sure. So is a drunken bloke groping a drunken woman at a party.
However, out in The Real World, there are so many other, more important "corporate evils" to be concerned about - in particular ones where actual people suffer actual harm due to events outside of their control - that the SlashGeek call to arms over "Micro$oft is teh 3v1l !" doesn't even pass the laugh test.
Well, if you've never read a book about logic, maybe it does.
Maybe if you've never read a book about logic, "I've never seen Windows win in a head-to-head competition" carries more weight than "I've seen numbers that indicate it might handle CPU sharing as well or better" as well.
Simple fact is that Windows has been tuned and developed for much bigger systems than OS X runs on, for much longer. Its current implementation has distinct technical superiorities to OS X's (eg: more fine grained kernel locking, more mature Async I/O) that result in it scaling better.
It applies to all cursor controls I've used mice, tablets, trackballs, pads, joysticks, even. If you haven't run into this, I can't imagine you've ever done graphics work.
I haven't, nor did I ever say that I did. I *have*, however, interacted with people who do - and none of them use the mouse for "drawing lines".
Yeah everyone is making it up because we're all part of a conspiracy... that's it.
Oh, grow up.
I didn't say anything about a "conspiracy", I said Mac users have been complaining about Windows's mouse tracking since day dot and *in my opinion* it's because of the different acceleration algorithms the two OSes use. I can relate, since the mouse acceleration in OS X (and Classic) used to annoy the hell out of me as being too slow. Now I just adjust my technique to take advantage of the different acceleration curves, but it's still noticably different for a good hour of use after moving from Windows to OS X.
A lot of Windows users get shoved in front of a mac, use it for a while, and then realize the problem they learned to live with was not an inherent flaw in all devices but a Windows bug that has never been fixed. Or it's all some form of mass hallucination, you pick the more likely one.
I use Macs, Windows, Linux and Solaris quite frequently. I have never noticed _any_ "mouse tracking bug" in Windows on the scale you are talking about, compared to other platforms. I *have* noticed that they all have different pointer tracking characteristics, and I *have* noticed people from one platform complain about that - but that's just a matter of adjusting to something different.
I didn't know you were referring specifically to UI responsiveness, but actually I remember seeing something (at arstechnica I think) a while back that did just that on a macbook running Windows then OS X.
Well, I'm not sure how else you could interpret the term "sluggish" in the context of multitasking, but that's not really important.
Ars are good for that sort of thing, but Siracusa tends to only compare to earlier versions of MacOS [Classic], not Windows. Some results from Google, like this one, do not do a lot for supporting any assertion that OS X is faster (although I believe there's recently been some patches that specifically address WoW performance).
Hahahaha! That is why it is better for power users. You see for larger numbers of windows, picking the application then the windows results in many fewer key presses and UI interactions. Think 10 applications each with 10 windows. To get from app 1 window 1, to app 5 window 5, with windows you have 50 key presses.
Or one, if it was the last window you used...
To get there in OS X you have 10 key presses (5 to get to the right app, then 5 to get to the right window). This only applies to larger numbers of windows and has a lower learnability than the Windows version. For power users who do it a lot, changing the key you press is not really any slower but pushing it an addition 40 times is.
Changing the key is important because it's not just changing the key, it's changing the mental model of what you're doing from "finding the right application" to "finding the right window". In Windows, the cognitive load is lower because you only have to be concerned about "finding the right window" - the applicatio
My father once told me: "you cannot be neutral between good and evil".
Absolute judgements are rarely accurate.
Not to mention, there's a vast, vast difference between "good and evil" and "something and Microsoft". On the scale of corporate bad behaviour, wouldn't even make it past the cutoff for the bottom quarter.
Anyone who calls Microsoft "evil" without their tongue firmly buried in cheek, *seriously* needs to get of their mother's basement, take a reality check and get some perspective on The Real World.
Maybe they're thinking that Apple didn't have a bad idea with OS X ... but where Apple went with Mach and a BSD userland, Microsoft could take a Linux kernel and then wrap an interface and a Windows API compatibility layer around it.
To what benefit ? There's little - if any - technical advantages the Linux kernel has over the Windows kernel. Indeed, many of the things that are "new" in the Linux kernel, the Windows kernel has had for a very long time, if not forever.
Of all the things Microsoft need to "fix" in Windows, the kernel is a long, long, long way down the list.
What about in ten years, when said Slashdot Geek has long graduated from university and gets promoted to a position where they can start to make purchasing decisions? (Apologies to those with 4-digit UIDs, who may already be old enough for this to have happened.)
By that time, most of them will no longer *be* the Slashdot Geek.
Ok, then you agree that there is something unique and interesting with the finder. Great.
I agree that they're useful. They're not "unique and interesting" - firstly because they're not unique and secondly because they've been around for nearly a decade now, putting them well and truly out of the "interesting" phase.
But nobody said anything about it being new. No argument there.
The OP was specifically referring to the timeframe while Microsoft has been working on Vista. Which - at worst - starts around late 2001. Specifically, he said Apple "introduced a new method of file browsing (love it or hate it, Finder is unique and interesting)".
(There were other errors in his post as well - eg. about OS X being "totally new", but the Finder one was the most conspicuous.)
Now as to replacing it, that is what spotlight will hope to be doing. Finder is certainly better than spotlight at this stage of the game. Definitely better than window explorer. At least I can drop to a real prompt and make any hierarchy I want with hard and soft links.
Comparing Spotlight to Explorer (or Finder, for that matter) is like comparing Word to Excel - nonsensical. While there's a slight overlap in functionality, they're meant for different things.
Services are no good if there is user interaction. I'm talking about applications. Tray icons are a hack and confuse most users to no end. They don't behave as users expect and don't have any way to provide the normal menus. Small windows are the common hack used to get around this problem. I've seen plenty of programs that have small, useless windows that take up space but are otherwise useless. So yeah, there are ways to work around the problem, but they are just nasty UI hacks.
I must admit I'm struggling to see what type of application you're talking about that needs user interaction, but can't justify a small window to put it in. It's not like Windows or OS X aren't already rife with these sort of "applets", nor that they violate existing UI guidelines.
I wish that were the case, but it just isn't so.
Well, I wish OS X was as fast and responsive to use as Windows, so I could use it full time without becoming frustrated - but it just isn't so.
I've seen numbers that indicate it might handle CPU sharing as well or better, but in the real world with disk and memory and other bottlenecks I've never seen Windows win in a head-to-head competition.
What "competitions" are you talking about ? Windows runs on much, much bigger systems than any version of OS X does and has been doing it for a lot longer. That alone, suggests it is more mature at utilising hardware resources and scaling to them. Heck, it's only in the latest versions that OS X has _started_ to implement fine-grained locking in the kernel (and is not even close to completing it). From a scalability perspective, OS X would barely be on par with Windows 2000.
One of the reasons macs are so popular among graphic professionals is if you're drawing a line on an OS X system it is unlikely that a background application will ever take enough resources to make the mouse input be ignored by the system or not displayed with the cursor. This is a common occurrence on Windows systems. Maybe this will be fixed with Vista as I have not tested it extensively yet, but for WinXP, Win2K and earlier this is certainly the case.
Can't say I've ever experienced it, but then again I rarely draw lines with the mouse. Nor do most graphics professionals I've ever met, either, they use graphics tablets.
Mac users have been complaining about Windows's mouse tracking for a couple of decades now. Personally, I think it's just the different methods in mouse acceleration rather than any real problem.
Your first statement is not supported by the benchmarks I've seen.
What cross-platform benchmark measures UI responsiveness ?
Both OS X and Windows are faster at different tasks with different applications. Now that both run on the same hardware, this is even more apparent. As for switching applications, it is better right now on OS X than it is on WinXP and historically, it has been a draw depending upon your workflow and experience.
I find Windows far more efficient for my usage pattern, which is lots of separate windows being switched between relatively frequently.
For people using the mouse as their switching mechanism, OS X is a clear win. For people using the keyboard, OS X was behind up until a couple of years ago when they extended the keyboarding and now, it is maybe a bit behind for novices and a bit ahead for power users.
Er, no. Task switching in OS X via the keyboard is demonstrably slower because it requires multiple key combinations to do so. You're either using Cmd+` to move between two windows in the same application (same as Windows) or you have to use Cmd+Tab to move to the right application, then potentially Cmd+` to move to the right window. In Windows, it's a single key combination.
OS X's UI is application-centric. It's designed with the assumption that users will spend relatively large amounts of time using a single application, possibly interacting with multiple documents within that application, and
I also doubt that Oracle would be stupid enough to limit support for their DB to Unbreakable linux. Their DB is their bread and butter, if they drop support for any OS that they curently support (say RHEL), there will be customer attrition. Why would they want to do loose Database customers for the sake of a product (Unbreakable Linux) that they will not make as much money on?
For Oracle customers, the only point of interest is Oracle. It doesn't matter if it's running on Linux, Solaris, Windows or a room full of monkeys banging rocks together - as long as the millions of dollars they have sunk into their environment, plus the millions of dollars in generates in revenue, that relies on Oracle to continue running, does so.
There are very, very few Oracle customers who are in any position to move away from Oracle. If Oracle says their DB will only be Supported on "Oracle Linux", then "Oracle Linux" is what's going on the DB servers.
Ideally, every software should be implemented in a way to make such changes, if not trivial, at least possible, regarding time and cost constraints.
Nice in theory, but it inherently either prohibits any company from implementing software that does $TASK better than anyone else, because they are constrained by having to retain compatibility with all the software that doesn't do $TASK very well, if at all, or it ignores the possibility of any $TASK that isn't designed and specified by a committee.
Or, to put it another way, for the result you want, all "innovation" - especially from individuals/small groups - has to cease, *OR* every developer has to work in collaboration with every other developer and sacrifice their ability to create a competitive advantage.
In other words, it ain't going to happen. Ever. As you note, the real world brutally intrudes on academic theory at the point where real money gets involved.
I think spring loaded folders are unique and interesting, and certainly useful...
They certainly are - but they date to 1997 when MacOS 8.0 was released (and even earlier, IIRC, from third-party addons).
On Windows, this type of application would confuse the crap out of people.
In Windows, this type of application tends to be either run in a service, minimise to a tray icon or simply display a small window (and is extremely uncommon, in any event).
Further, I often want to close all the documents in use by an application, but not quit the application. The reason for this is also simple, I don't want to wait for it to load again multitasking us good enough that having it open does not slow down other applications. I think this has also been an issue related to Windows since on Windows users have trained themselves to quit applications not in use, since the multitasking does not handle programs sitting idle in the background very well.
IME Windows multitasks much better than OS X does. Firstly because OS X is so sluggish to begin with - and adding more running apps just makes it worse - that interacting with multiple tasks is discouraged, and secondly because the task-switching UI - particularly historically - isn't as good. Expose certainly added a dramatic improvement on previous versions (and cleverly hides some of the sluggishness with flashy graphics), but it becomes difficult to use with numbers of windows in the double digits, especially if they are similar (eg: terminals) or screen resolution is relatively low (a problem with most Macs).
Interestingly (with regards to your comment) while Microsoft was doing UI research for Vista, they found that Windows users tend to "multitask" - that is, run more stuff at once - more than Mac users did. They also tended to switch between tasks more frequently. This was outlined in the Vista GUI blog that was linked from Slashdot a month or two ago. It also mirrors my experience, both as a user and observer and, in my opinion and experience, it's because Windows simply handles multitasking better and has done so for much longer. I certainly know that every Mac I've ever used has choked up under the sort of multitasking load and usage pattern even my old dual P3 could handle (running Windows or Linux). My 1Ghz iBook can't even handle as much stuff simultaneously as a Compaq 500Mhz P3 notebook, despite having ~30% more RAM and easily twice the raw processing power (although, obviously, individual processing-intesive tasks are much faster).
It pisses me off when I'm on a Windows box and I close a file only to have reopen the program that just quit because I want to open another file using that same program.
If you're immediately reopening an application it should be essentially instantaneous, since all the relevant data will still be cached - unless your machine is memory starved.
This is a common task that Windows application developers hack their way around by creating managing Windows that just sit around as placeholders with nothing in them, or useless features in them that only duplicate functionality of the menus.
What applications do this ? I can't think of any off the top of my head, so I highly doubt it's a "common task that Windows application developers hack their way around".
There is one difference and that is the dock. If you close a window the user will notice the program is still running pretty quickly, since the dock icon remains and is marked as running. Most novice mac users use the dock to start and stop applications. It makes a lot more logical sense, in my opinion, than tying the action to having open windows.
The Dock's indication of what is and isn't a running program is, IMHO, far too subtle. Certainly many novice users I know (including my parents) have trouble with it. Of course, IMHO the Dock as a whole is a UI train wreck, and a significant step backwards from the equivalent UI in MacOS Classic, and especially Windows. Certainly looks cool, though.
Every screenshot I've seen of Aero looks remarkably...Aqua-ish.
Can you think of a way to implement various things like transparency and fancy effects today that *wouldn't* look "Aqua-ish" ? Follow-on competitors are always going to "look like" the first product to market they're competing with to some degree.
Really, if you're comparing just the two of them, they're not especially alike. The way they use transparency, fading and other effects (like "sheets" rolling in and out of windows, or "Flip3D") is all quite different.
I've been using Macs for a very long time, and Vista daily for a couple of months now, and the two have vastly more differences than they do similarities - and most of those similarities are the kinds of things that can appear even with two completely independent implementations - just like MacOS and Windows have had, well, pretty much forever.
Not in the details, but I can't help thinking that someone at Microsoft took a look at Aqua, and decided that it was probably time to overhaul Windows' interface as well; not to mention doing the same sort of graphics-card offloading that Apple did with Quartz Extreme.
The Windows GUI has been getting the same kind of "overhaul" it gets in Vista with pretty much every release since Windows 95 (which was the last _major_ change). It's reasonably safe to say even without the pressure of OS X, it would have gotten a similar update to the one it actually did - if only because 99% of people (even on Slashdot) don't look any further than that to see what's "new and improved".
Moving the display system onto the video card is just a natural and obvious progression. Microsoft had been talking about doing it since before Apple did, although OS X obviously beat them to the market with an actual implementation. Saying they were "copying Apple" in this regard is specious, at best - about as accurate as saying Apple were "copying Microsoft" by implementing an OS with pre-emptive multitasking and protected memory.
Introduced a new method of file browsing (love it or hate it, Finder is unique and interesting)
There is nothing "unique and interesting" about Finder. In two of its three modes, it's just a (less functional in some ways) copy of the MacOS Classic Finder. In the third, it's a copy of the NeXT file browser.
(And they all suck, IMHO, but that's not reall the question here.)
Another of Apple's little fits of genius was to make the laptop automagically write the RAM/states to disk - essentialy switch to a hibernated state - when battery power drops to a dangerous level. Waking then takes a few extra seconds, but is still much faster than a reboot - and of course the user state is precisely as you left it. So closing the lid on an iBook effectively lasts forever, because Apple was smart enough to make the distinction between Sleep and Hibernate transparent to the user.
Which iBook ? My 1Ghz iBook doesn't do this - if the battery goes flat when it is sleeping, it does not resume, it boots to a fresh desktop.
Ah! So X is only half of a GUI. Gnome and/or KDE is the other half. That makes sense... kinda. Although it does raise the question "If X has been around for such a long time providing the bottom half of a GUI, why hasn't there been an equally seasoned top half?"
Becuase the only really useful feature unix nerds see in X is the higher resolution allowing them to fit more xterms on the screen...
I've always found it ridiculous how Mac users don't like running cross-platform applications under X.
That's because you're not a Mac user.
You simply cannot spend 2 years cranking out a dot level upgrade to an OS and expect thundering success. Because in the final analysis there isn't much new to Vista
Vista is a solid 5.x ->6.x revision. The changes - particularly in the lower levels - are significant, on par with the updates Apple did to NeXT to get to OS X. That most people do not bother looking past the relatively minor GUI update does not change this.
Extrapolate the relative improvements in Windows and KDE and Linux to 2010.
You assume the rate of improvements will remain constant. A shaky assumption at best.
I'm not a MS fan, or really anti-MS - (read my other posts) but Microsoft are stuck here, they cant simply adopt a radical new operating system (even with virtualisation) and throw out all legacy support. They have gone as far as they can with Vista (which really isn't a re-write, its a mash up and an attempt to address some of their major problems).
Microsoft's problem is not their codebase or technology, it's their development process.
With legacy support they will continue to have the same security problems that have plagued them since the inception of windows.
Any platform with the marketshare and userbase Windows does will have the same problems Windows does.
The reasoning is simple, if Microsoft adopted a *nix like kernel and re-wrote everything then they would have an OS with little or no software support, little or no Hardware support, and they would find themselves competing directly with Linux / Solaris / BSD etc...without the benefit of already having a huge installed user base with a clear upgrade path. There would be one additional exception, their offering would not be free (unless they would simultaneously go down the software as a service route and make their OS free.. which throws up even more issues see My other post)
They'd also be taking a technological step backwards or, at best, sideways.
The Windows NT kernel is about the *last* thing Microsoft needs to be changing.
Sorry, but if Apple, with way less resources, was able to completely change their OS, change platform and still provide a migration path for developers, why microsoft, with vastly superior resources and cash, would not?
Because firstly, and most importantly, Apple has fewer scruples about breaking/obseleting old hardware and software.
Secondly, because Apple's user base is an order of magnitude smaller.
M$ knows this, if they change the model T-Ford like kernel, applications will break.
Uh, if the Windows NT kernel is a "Model T Ford", what does that make Linux ? A horse-drawn cart ?