Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

So what do you think about Macs/Apple OS?

  • They are superb and could not be better

    Votes: 305 22.9%
  • They're good but have a few niggles

    Votes: 879 65.9%
  • For everything I like there's something I don't like

    Votes: 106 8.0%
  • I prefer Microsoft PCs

    Votes: 43 3.2%

  • Total voters
    1,333
Status
Not open for further replies.
Windows Vista.

Leopard has it too, just not accessible to the front-end user yet, (heard rumours a while back of being switched on between 10.5.3 and 10.5.6, but no sign as of yet).

Honestly once Resolution Independence is available to tweak it will look amazing. Examples are looking at the Dock resizing and Cover Flow in Finder. Seemless and scale perfectly. Windows Vista requires a full restart for any change to take effect and even then most changes just look like you've dropped the screen resolution. Apple have partly already done and will almost certainly do it better.
 
What I hate the most is how a stupid program making heavy use of the HDD can render the OS completely useless for upto minutes at a time. But then again I have not yet seen a OS that does no suffer from this. I cant believe with all the hardware advancements of the last many decades we still have to suffer the OS being hijacked by a single program. Write a proper disk/CPU scheduler with proper priorities for gods sake.

Also most 3rd party programs for OS X suck, both in terms of performance and the tendency to crash incredibly often when performing the simplest of tasks. QT player is very badly written, in term of performance and stability when it comes to playing anything other than what you can get off of iTunes. Also, too often crashing programs somehow manage to take the OS with them in their death spasms, in theory something like that should be impossible, yet it happens.

File associations are also pretty retarded sometimes where OS X refuses to open a file type with the program I tell it to use.

I guess Apple is finally rolling up their sleeves and are going to be doing something about the resource gluttony of OS X. But for now add the fact that any OS X programs eat up unfair amounts of memory so those 2 gigs are gone in a flash and then the swapping and beachball begin.

In fact any time I see the beachball is a moment where I hate my Mac.
 
What I hate the most is how a stupid program making heavy use of the HDD can render the OS completely useless for upto minutes at a time. But then again I have not yet seen a OS that does no suffer from this. I cant believe with all the hardware advancements of the last many decades we still have to suffer the OS being hijacked by a single program. Write a proper disk/CPU scheduler with proper priorities for gods sake.

Also most 3rd party programs for OS X suck, both in terms of performance and the tendency to crash incredibly often when performing the simplest of tasks. QT player is very badly written, in term of performance and stability when it comes to playing anything other than what you can get off of iTunes. Also, too often crashing programs somehow manage to take the OS with them in their death spasms, in theory something like that should be impossible, yet it happens.

File associations are also pretty retarded sometimes where OS X refuses to open a file type with the program I tell it to use.

I guess Apple is finally rolling up their sleeves and are going to be doing something about the resource gluttony of OS X. But for now add the fact that any OS X programs eat up unfair amounts of memory so those 2 gigs are gone in a flash and then the swapping and beachball begin.

In fact any time I see the beachball is a moment where I hate my Mac.

I have none of your issues, besides the first one...
 
Leopard has it too, just not accessible to the front-end user yet, (heard rumours a while back of being switched on between 10.5.3 and 10.5.6, but no sign as of yet).

Honestly once Resolution Independence is available to tweak it will look amazing. Examples are looking at the Dock resizing and Cover Flow in Finder. Seemless and scale perfectly. Windows Vista requires a full restart for any change to take effect and even then most changes just look like you've dropped the screen resolution. Apple have partly already done and will almost certainly do it better.

Steve,

That would be great news, but I have searched again, and again; and I have found nothing about Apple adding Resolution Independence. Sad that Vista has it, and Apple can't seem to implement . . . . or even talk about it.

L
 
What I hate the most is how a stupid program making heavy use of the HDD can render the OS completely useless for upto minutes at a time. But then again I have not yet seen a OS that does no suffer from this. I cant believe with all the hardware advancements of the last many decades we still have to suffer the OS being hijacked by a single program. Write a proper disk/CPU scheduler with proper priorities for gods sake.

There is no avoiding this other than increasing the speed of the device. It's not the OS' fault. It's a hardware problem. The OS can't magically prioritize access to a bottlenecked resource. Breaking up IO to a slow interface is only going to make things worse. You need to move large blocks to be efficient and that takes time on a slow device. Moving smaller blocks of data to allow for "proper disk scheduling" will only make things worse.

This is one of the reasons I consistently rail against higher and higher cpu rates while bus, memory and disk speeds remain fairly constant.

A buddy of mine and I both bought linux servers at the same time. Both servers were from the same manufacturer, and were the same base platform. He went with the best cpu's. I went with the fastest bus, memory and 15k rpm SCSI disks, but the cheapest cpu's. He spent more money and my server beat his in every real world application test we compared them with.
 
Steve,

That would be great news, but I have searched again, and again; and I have found nothing about Apple adding Resolution Independence. Sad that Vista has it, and Apple can't seem to implement . . . . or even talk about it.

L

Install the Developer Tools and then do a Spotlight search for Quartz Debug. In the Tools menu select "Show User Interface Resolution". Set the resolution to anything greater than 1 and then launch any app. The app will now be drawn by Core UI and be resolution independent.

The foundation is there. But not all the UI elements are finished. Round textured buttons default back to their Aqua Tiger versions for example.

Here is TextEdit with a UI resolution of 1.5:

picture13e9c5.png


picture2cdc4d.png


Compared to TextEdit with a UI resolution of 1:

picture300c92.png


picture4ddfa9.png


Eventually this will be accessible by the user and will work on the fly like the Dock does when you switch the resolution.

An example of how the UI elements are quite unfinished at the moment:
picture50ad7a.png
 
The delete key doesn't delete anything. Instead one has to either drag it to the trash manually or pick "move to trash" from the contextual menu.

try holding the Apple :apple: button then deleting..
i find nothing wrong with Mac's- Probably the best thing i've ever brought..
 
try holding the Apple :apple: button then deleting..
i find nothing wrong with Mac's- Probably the best thing i've ever brought..

Command-backspace every time for me. Tip, if you don't know the keyboard shortcut, go to the menu and it's listed next to it. It's a little extra safeguard if you're accidently deleting items when Finder is selected and you didn't realise.

One annoyance on the back of that though is when you do move the wrong item/too many items to the trash, there is no way to restore them back other than moving them back to the folder, (unless I've missed something). Windows does give you the option to restore back to original location.
 
In fact any time I see the beachball is a moment where I hate my Mac.

This, and I see it a heck of a lot on my Powerbook. But then, it's old. And I also hate my PC every time I see the hourglass (or spinny circle thing these days) and I seem to see it a lot more there.

The thing that's loathsome about beachballs, spinners and hourglasses isn't so much the time spent looking at them, but the fact that you don't know what's going on. For humans, it's very uncomfortable waiting for something when you don't know what's happening and don't have any indication of its progress.

If you know that the new exhaust is going to take 2 hours to fit... that's great! Off you toddle to the nearest restaurant or shopping centre, safe in the knowledge that you'll come back and it'll be done. But if all you see is your car being taken into a garage and then you have to stand by a closed door, it'd drive you nuts after less than 5 minutes. 2 hours later you'd have no hair left, and would be cursing them for not telling you that it'd take 2 hours.

This is why I think non-progressive "busy" indicators should be an absolute last resort in UI design. If the disk is busy indexing a directory then say so - underneath that beachball put a little progress bar and description like: "Indexing Home\Pron [----- ] (30 seconds remaining)". People would be so much happier.

Or even a little Growl type message that pops up in the background somewhere saying "OS X is currently running its daily maintenance routines. Your computer may slow down for approximately 10 minutes."
 
- The sleep light cannot be disabled. I like not shutting down my Power Book but, in a dark bedroom, that light gets annoying.

At least the Mac Book Pro sleep light is not as harsh as the Powerbook. I still have to put a sock over it every time I go to bed.
 
Also most 3rd party programs for OS X suck, both in terms of performance and the tendency to crash incredibly often when performing the simplest of tasks. QT player is very badly written, in term of performance and stability when it comes to playing anything other than what you can get off of iTunes. Also, too often crashing programs somehow manage to take the OS with them in their death spasms, in theory something like that should be impossible, yet it happens.

File associations are also pretty retarded sometimes where OS X refuses to open a file type with the program I tell it to use.

I guess Apple is finally rolling up their sleeves and are going to be doing something about the resource gluttony of OS X. But for now add the fact that any OS X programs eat up unfair amounts of memory so those 2 gigs are gone in a flash and then the swapping and beachball begin.

In fact any time I see the beachball is a moment where I hate my Mac.

What the heck do you do with your computer on a daily basis? Sounds like you're hard on your system in many ways. You're right about the system slowing down on certain programs and that's typical of any computer. I have major doubt that you've tried the many thousands of 3rd party software for Mac OS X so think twice before you say "most 3rd party programs for OS X suck". I actually prefer to use 3rd party programs rather than Apple's, they have less bloat than Apple's and updates are generally more satisfying.
 
There is no avoiding this other than increasing the speed of the device. It's not the OS' fault. It's a hardware problem. The OS can't magically prioritize access to a bottlenecked resource. Breaking up IO to a slow interface is only going to make things worse. You need to move large blocks to be efficient and that takes time on a slow device. Moving smaller blocks of data to allow for "proper disk scheduling" will only make things worse.

No, there is plenty of avoiding it, it is called writing a bleeding smart as a fox disk scheduler. A clever disk scheduler should at the very least under no circumstances allow any process to spam the disk I/O queue. It should be written in a way where it detects that a process is in the process of doing a crap load of reading/writing to the disk and then allow the read/write requests of other processes to jump ahead of the queue, specially, if they are related to system processes, ESPECIALLY!!!!

Combine that with whole hell of a lot more clever rules for how to priorities simultaneous read/write request from different processes and it should be very possible to create an OS where programs that rape the disk I/O can still slow the system down, but not stall it so it becomes so useless that you have no choice to walk away and do something else in the meantime. It is 2008, the concept of multitasking was invented decades ago and yet I have to often suffer through a program giving OS X a joyride in ye old timemachine, back to the time of a single task operating systems.

This is one of the reasons I consistently rail against higher and higher cpu rates while bus, memory and disk speeds remain fairly constant.

Yes, exactly, for the life of me I cannot understand how incredibly little development or focus there has been dedicated to this mother of all bottlenecks in computers. I swear to ye old gods, whenever I hear my hdd having a seizure as the beachball does its hypnotic dance across my screen trying to add a little cheer to the insufferable business of my Mac being rendered useless I am overcome with a strong urge of hammering a screwdriver through my Mac exactly where the HDD is while laughing manically. One of these days I am going to act on it, blast be that it wont solve anything, it will make me feel better dammit.

What the heck do you do with your computer on a daily basis? Sounds like you're hard on your system in many ways. You're right about the system slowing down on certain programs and that's typical of any computer. I have major doubt that you've tried the many thousands of 3rd party software for Mac OS X so think twice before you say "most 3rd party programs for OS X suck". I actually prefer to use 3rd party programs rather than Apple's, they have less bloat than Apple's and updates are generally more satisfying.

I do only 3 things on my and mostly only these 3 things. I download videos from the net, mostly from News servers which then need to be Par checked and unrar/unzipped. I watch same said videos, and lastly I surf the web. I am neither happy with how OS X performs during any one of these 3 tasks nor am I happy about the programs available to do them.

I have tried all available programs that can do unrar, and none of them can even being to hold a candle to winrar. Extremely few of the unrar programs can handle all types of rar file formats and the select few that can are very amateurish in execution. All of them are glitchy and often during unrar land on their asses so hard that I have to use Force Quit to get them off of it.

The only semi descent par check program for OS X is MacPar Deluxe and lets face it, it sucks, very much. Surprisingly it does not have a problem with crashing, but what it tries to make up for in stability it more than fails in performance. If bad performance was a sin then Mac Par Deluxe would be Judas, or an equivalent backstabbing bastard of you favorite religion.

More than any other program MPD renders my Mac completely useless, it makes spamming the I/O queue into an artform and does with a skill unseen by my eyes, on either sides of the fence, and still somehow finds energy to punch both of the proud Intel Core 2 Duo cores of my 2.2 ghz MacBook in the balls. I am windows refugee, so all the programs I use on my Mac I used some other version of on Windows. MPD's windows alter ego is called QuickPar and at least in comparison it is friggin awesome.

When I start the parity check on OS X with MPD all programs become useless, specially Safari which just beachballs for the upto 10-20 min it takes for the parity check to finish. If, god forbid, an invalid checksum is detected and MPD attempts to fix it then go for a walk and leave the windows open. My already overheating MBP becomes so hot that it might as well be a frying pan and forget about using the Mac unless you want to use it as one of them fancy beachball renderers.

The video watching part? Well forget all about Quicktime, because if by some miracle it can actually open the video then it is surely to either stutter or crash. I only watch HD movies, if you are curious, mostly x264 or mpeg2 TS encoded with AC3/DTS sound and often in a MKV container. Quicktime is designed to playback at the very most SD resolution mpeg4 or simple profile H264 with 2 channel AAC audio, and pretty much nothing else. That is all good and well from Apple's point for view, it can play back anything from iTunes reasonably but it is useless for anything else.

I use either VLC or OSXMBC for video playback and the biggest problem of both is stability, specially VLC. Even a gentle gust of wind can sometimes bring down VLC, but at least it has Quicktime beat by lightyears in terms of performance and video codec/container support. OSXMBC has them both beat but it too is plagued by horrible stability and has managed to kill OS X a few times too, something Quicktime is also an expert in.

Other than that I have to say unlike most Apple written programs Safari is pretty terrible what performance and stability is concerned. I have a tendency to have a lot of windows and tabs open at the same time and Safari does not like that, no sir, not the least bit. As soon as the system becomes heavily loaded Safari throws in the towel. It crashes often and it does not seem like it is very good with threading, where I can work in one Safari window or tab as another is doing some heavy loading. If one is overburdened then they all become completely unresponsive and the beachball does its dance.

And aye, all them HD movies are bootleg, just ye try and stop me, I'll run ye though with me pirate saber.
 
No, there is plenty of avoiding it, it is called writing a bleeding smart as a fox disk scheduler. A clever disk scheduler should at the very least under no circumstances allow any process to spam the disk I/O queue. It should be written in a way where it detects that a process is in the process of doing a crap load of reading/writing to the disk and then allow the read/write requests of other processes to jump ahead of the queue, specially, if they are related to system processes, ESPECIALLY!!!!

You are incorrectly applying cpu multitasking concepts to an I/O device. You cannot divvy out disk I/O to applications like you do with the cpu without creating even longer delays due to head thrashing and inefficient use of I/O buffering.
 
You are incorrectly applying cpu multitasking concepts to an I/O device. You cannot divvy out disk I/O to applications like you do with the cpu without creating even longer delays due to head thrashing and inefficient use of I/O buffering.

I am a programmer, I cannot believe the only solution is to allow programs to spam the I/O queue.
 
I am a programmer, I cannot believe the only solution is to allow programs to spam the I/O queue.

That's why you aren't engineering I/O interfaces. ;)

Obviously the OS doesn't allow a free-for-all to occur, but there is no simple solution like you are saying..
 
How can you be so sure?

Because I've educated myself in how I/O devices work and are programmed. And because it makes a pretty hilarious conspiracy theory to assume that every OS manufacturer out there refuses to spend a few hours of programming to avoid one of the nastiest things that can happen to their users so they can collectively laugh at all the hourglasses and beach balls they've thrust upon their users for no good reason other than evil pleasure.
 
Because I've educated myself in how I/O devices work and are programmed. And because it makes a pretty hilarious conspiracy theory to assume that every OS manufacturer out there refuses to spend a few hours of programming to avoid one of the nastiest things that can happen to their users so they can collectively laugh at all the hourglasses and beach balls they've thrust upon their users for no good reason other than evil pleasure.

I always liked to pretend it was done not due to malcontent but rather due to crippling incompetence, you know, the reason behind the million other things wrong with any OS.

And I dont see it as an hardware problem at all, my solution is a software solution. The OS receives and processes all disk I/O requests from programs, it should be able to implement some sort of software queue with scheduling for these requests before they are committed to the hardware.

I admit I know nothing about the hardware side of things, but simply seen I imagine it as a device that does what you tell it to do in the order you tell it to do.

The OS tells the I/O what to do, so it is free to implement a software scheduler with a tiny bit more fortitude than a rock. Should be possible to insert some higher priority disk writes or reads ahead of a queue if it has been spammed by a program that is not likely to stop anytime soon.

Sure it might slow down the process of the spamming program but at least it will make other programs more responsive. Thrashing can be avoided through smart coding. Like analyzing what a program has been doing while it has been running or store a record of its past behaviors or even just looking at the current queue. Besides, how can the current situation be considered any better than thrashing?

You wouldn't need smaller read/write buffers either, I am not talking about milisecond problems here, but about programs becoming unresponsive for minutes, even hours at a time. If there is absolutely no possible way to avoid that then I think all software and hardware engineers in the world should just hang themselves out of shame for their own incompetence, I'll tip over my chair as soon as I see all the people at Apple, M$ and all the Linux/Unix folks have stopped twitching.
 
I always liked to pretend it was done not due to malcontent but rather due to crippling incompetence, you know, the reason behind the million other things wrong with any OS.

And I dont see it as an hardware problem at all, my solution is a software solution. The OS receives and processes all disk I/O requests from programs, it should be able to implement some sort of software queue with scheduling for these requests before they are committed to the hardware.

I admit I know nothing about the hardware side of things, but simply seen I imagine it as a device that does what you tell it to do in the order you tell it to do.

The OS tells the I/O what to do, so it is free to implement a software scheduler with a tiny bit more fortitude than a rock. Should be possible to insert some higher priority disk writes or reads ahead of a queue if it has been spammed by a program that is not likely to stop anytime soon.

Sure it might slow down the process of the spamming program but at least it will make other programs more responsive. Thrashing can be avoided through smart coding. Like analyzing what a program has been doing while it has been running or store a record of its past behaviors or even just looking at the current queue. Besides, how can the current situation be considered any better than thrashing?

You wouldn't need smaller read/write buffers either, I am not talking about milisecond problems here, but about programs becoming unresponsive for minutes, even hours at a time. If there is absolutely no possible way to avoid that then I think all software and hardware engineers in the world should just hang themselves out of shame for their own incompetence, I'll tip over my chair as soon as I see all the people at Apple, M$ and all the Linux/Unix folks have stopped twitching.

Do some reading about head thrashing and I/O buffering on disk devices.

As painful as it may seem, it's generally better to allow disk hogs to read large blocks rather than to cut them short repeatedly to serve smaller requests. I/O buses perform much better with large chunks of data, buffers are rendered nearly useless if you don't read large chunks, and constantly making the disk heads move more times than is necessary (which would be mandatory to allow "premptive disk IO") is utterly crippling to disk I/O speeds.
 
Do some reading about head thrashing and I/O buffering on disk devices.

As painful as it may seem, it's generally better to allow disk hogs to read large blocks rather than to cut them short repeatedly to serve smaller requests. I/O buses perform much better with large chunks of data, buffers are rendered nearly useless if you don't read large chunks, and constants making the disk heads move more times than is necessary (which would be mandatory to allow "premptive disk IO") is utterly crippling to disk I/O speeds.

I know about all this, but how big are those chunks you are talking about? Kilobytes, megabytes or hundreds of megabytes?

I imagine you are referring to megabytes, because most files are fragmented around the harddisk anyway so you cannot expect to be able to read hundreds of megabytes without having to move that needle around some anyway. So lets imagine a program that wants to read a gigabyte, process it and write it down, kind of like a unrar program.

This program would read a chunk of maybe a megabyte into a buffer, process it and write it back. Thanks to CPU speeds rocketing upwards for every year these last 3 decades the CPU will finish its job so fast that the time it takes is negligable. So the program is essentsially just issuing 1 megabye large read and write commands in succession.

Now, the OS that is executing the code of this unrar program is pretty much useless at this point. Try and fire up safari, the first thing it does is to try and read a bunch of data from the disk but it is stalled because its read request are sandwiched between the I/O spammer programs requests.

Now, the least I can expect an OS to do is to realize that for the last, oh, lets say 30 sec to a minute at the very least this stupid unrar program has been spamming the I/O unusually much and based on its past history it is likely to continue to do so. Safari on the other hand is just starting up and needs a few small reads, so, the OS ignores the requests coming from the unrar program for a while by putting them in a queue and focuses on Safari. Once that is done unrar gets to continue have its request executed to the HDD hardware.

The point is this is done in between one of the read/write requests coming from unrar. If the OS is smart, and I firmly believe it can be, then it can do all of this while at the same time avoiding head thrashing. The would have to collect data about different programs and their behaviours, hell it might pay for the programs to be able to tell the OS how they are expected to behave. It would require a whole new way of thinking.

Yes, potentially unrar might perform worse than if it was run in a more retarded OS but that would only be a bad thing in a OS dedicated to be a server. But I am talking about consumer OS's like OS X that should have been designed with multitasking in mind and as a priority. Let the server folks use Unix or some other lightweight efficient OS up to the task, I want a consumer OS that actually can multitask, not just claim to be able to.
 
I believe you are lacking the basic fundamentals of what is necessary for efficient IO. But I'm not going to continue the hijacking of this thread, if you want to start a new one, go for it.
 
I am not interested in efficient IO, I am interested in efficient multitasking even if it should come at the cost of IO efficiency. I have no more to say, no need for a new thread.
 
Nah, things don't work better in 9, sounds more like you are just being a cheapskate.:D

I have a friend who spent a fortune on ProTools plugins back in 2000. Then he had to re-buy them for OS X. It's kinda disappointing that he didn't have time to make money on the use of his new tools before they were outdated. It's not Apple's fault, it's no one's fault. But it is a good argument for Classic on Intel.

I notice in Leopard that column view doesn't preview .vob files like it did in panther and tiger. This is actually really getting on my nerves A LOT lately. If I was skilled I would write the quicklook plugin to do it, but I'm not a developer or code guy. As a professional videographer I'd say the flaw lies with Apple on this one.

To add to that chain of thought, the white borders around the pic files. DEATH to the u.i. team who did that!

The menus opacity in 10.5.3 bugs me. It bugs me way more that they didn't think to make the checkbox affect the menu bar and the menus.

Spotlight search results window. OMGWTF is wrong with Apple on this one. It makes me miss Tiger.

The "slow-mo" shift key makes it so you can't use it in combination for activating Exposé, Dashboard and Spaces.

The inability to right-click in a stack. Sometimes I want to choose "open with app" to select an app other than the default.

Trash management would be nice. I use my thumbstick a lot.

Can't navigate to "cancel" or whatever else without a mouse.

Leftover Aqua elements? Why didn't they finish up the unified look? Why does FCP 6 still have brushed metal?
^Probably because those elements need to be finished for the resolution independence switch. Too bad it wasn't done in time for Leopard's launch.

But seriously. The .vob thing is driving me up a wall this week. Does anyone know a fix???
 
Trash management would be nice. I use my thumbstick a lot.
Agreed.

Can't navigate to "cancel" or whatever else without a mouse.
You can. It's not obvious, but I believe you click the "Enable access for asssistive devices" under Universal Access in the System Preferences. Tab then toggles dialog buttons. Enter activates the default option while Space activates the encircled option from tabbing.

Windows does keyboard access better in most (but not all) ways, but OS X gives you many of the features.
 
Status
Not open for further replies.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.