There is a lot to complain about in OS X, but as a programmer I am not sure I would have picked the points he did.
Things like kqueue() support not being up to par with the other kqueue() providers, notably FreeBSD. kqueue() is considered so broken by libev for example that it just doesn't use it. The deprecation of OpenSSL (although I'm happy about that, but installing a secondary OpenSSL alongside the OS provided one can break things, so you have to install to a separate path and use include directives in your compile line, and or an alternate pkg-config path). The fact that a replacement libc++ now exists, but parts are still using the old stdlibc++ and the two can't be linked together if you use std::string for example, thus making it harder to use C++11 features. This is going to be an issue for a little bit, but hopefully that gets better. Outdated utilities or libraries make life more difficult, however a lot of that can be fixed with homebrew. Debugging USB issues is more difficult than it should be, this is especially an issue when attempting to figure out why a driver from a vendor crashes when you use certain USB to serial devices.
The new sandboxing requirement for apps while making everything a little more secure makes it more difficult to make certain types of apps ... slowly but surely new API's or permissions are being made available to fix those issues, or at least improve the pain points.
Do you have some links with information about kqueue being broken? I know that the python integration doesn't work well. We started using polling on OS X instead of watchdog's kqueue driver.
kqueue()'s support of process monitoring on OS X doesn't allow for tracking of new processes that are created, and automatically attaching to them.
There are also differences that make it more difficult to port various kqueue() features from one platform to another. I'm not sure how much of a difference that makes to you.
The documentation on OS X is also lacking, for example EVFILT_USER is supported on Darwin/XNU but it is not documented in the man page for kqueue.
Sorry, but you simply didn't read the docs when setting up backup of your critical data. It's explicit that Time Machine does not back up removable drives. Try: http://www.bombich.com/
EDIT: Apparently, this runs deeper. Time Machine mistakenly marks his internal drives as removables. However, he also states that he's done lots of system modding and that he has 10 internal drives. I don't think he's the target user TM was designed for. I'm not either. I have also had some TM problems and now use CCC as my primary backup, with TM as a convenience and secondary backup.
> Finder Refuses to Delete 'Backup Items'
Author states he uses OS X for serious work, and implies he has the technical wherewithal to evaluate OS X's stability and coding excellence. Apparently, he's not a programmer, or not one competent enough to use lsof and kill.
> Bottom line: Fusion as-is and as supplied by Apple is OK (maybe, hard to say for sure in multi-drive systems). Other than that, messing around carries some ugly risks.
Sorry, but if you mod your setup, then you should look in the mirror when you complain. For what you want to do there, you should be using diskutil corestorage on the command line. However, the diskutil GUI doesn't tell end users everything about corestorage yet, and I agree that this is bad. It contravenes expectations established by Disk Utility and OS X in the past -- without sufficiently warning users.
Yeah, that's bad, but I get the sense that he's doing some core storage stuff he shouldn't be doing. 10 internal drives? And he should know better than to use the Disk Utility GUI for serious corestorage operations.
What's the connection between a complaint about Time Machine's silent exclusion of critical data and Time Machine's missing support for removable drives?
Yo dawg I heard you like backups so I made a backup of your backup on your backup.
The idea here is that Time Machine is generally not useful on transient drives that come and go, thus for each time that you plug in the drive Time Machine would have to re-scan the entire drive looking for changes, to then back those changes up to yet another external drive. Time Machine is meant to protect the files on your computer, not those on external media.
Do I agree with that? No, I think I should be able to specify that Disk A should have a Time Machine backup on Disk B, and that Disk A and B are stationary to my desk, but most people don't need that. There is no requirement for that, people want to backup their main hard drive to an external so that in case something goes wrong their data is safe.
The authors complaint seems to stem from the fact that Time Machine can't find a drive to backup to (which BTW, Time Machine will warn you about every single day that it is unable to backup), probably because he spent an awful amount of time partitioning his drives in weird ways with weird names, and copying stuff back and forth thereby probably getting all drives tagged by Time Machine as not viable for a backup location.
--
Personally I haven't had any issues with Time Machine. I've got it setup over iSCSI to an OpenIndiana host and to a local 1TB external drive and all is well, haven't had any issues what so ever with it no backing up data, and restoring has been a breeze in the past.
Because he's been mucking around with core storage volumes and drive bay extenders to the point that certain of his internal drives look like they're on an external bus to the OS.
Apple optimizes for the common case, which he's way outside.
There's a lot of misinformation and half-truths in this article. To pick one example, the filesystem: yes, we're sadly not using ZFS but that's in all likelihood due to licensing restrictions, not because Apple thinks HFS+ doesn't need replacing. What Apple has done is introduce CoreStorage, a "storage virtualisation system" (see link below), which beyond allowing new features to be added now (e.g. full disk encryption), will presumably allow easier migration to a new filesystem in future. And this was added in Lion.
As a self-described power user, I've found 10.8 to be a nice upgrade on 10.7 and worlds better than 10.6. That's not to say there aren't issues -- there are many, as there are with any OS -- but Apple currently isn't showing any sign of slowing down, with many of the issues I've experienced having been fixed in 10.8, while any power user is quite capable of disabling new consumer-facing features that don't benefit them (e.g. Launchpad, iCloud).
Edit: reading further, there's a better example -- the author's criticism of Fusion Drives. I quote "Bottom line: Fusion as-is and as supplied by Apple is OK (maybe, hard to say for sure in multi-drive systems). Other than that, messing around carries some ugly risks."
Not only are Apple's Fusion drives a new technology (and it's a relatively pioneering one at that, at least in the consumer space -- they're not HDDs with an SSD cache, but rather SSD/HDD hybrids that intelligently move data from one to the other), thus disproving the point that Apple isn't innovating on OS X anymore, but his criticisms apply when you're making your own Fusion drive, rather than buying a computer with a Fusion drive installed. This is something unsupported right now, especially since Fusion drives have only existed for several months, so although of course it should be fixed I don't think Apple deserves too much criticism for this.
Second edit: this is perhaps controversial, but I think it comes down to the difference between a power user and a dangerous user. Power users who have to get work done will generally stick to supported system configurations, or will be damn careful (and with full knowledge of what might happen if things go wrong) if doing anything unsupported. I don't think any "power user" would have a custom Fusion drive setup and then complain about Apple when something goes wrong. That's not to say all of the criticisms in the article are wrong, because they're not, and some are quite serious. But some things are definitely disingenuous too.
I agree some things here aren't backed up, and there are half-truths. The guy says he's going to list bugs, then he really doesn't, he just bitches in bulletpoint form.
HOWEVER: He's totally right about everything behind all this. Apple is really fucking up OS X, and clearly it has no desire to keep any users that don't see their iMac as a giant iPhone.
My wife just got a brand new MacBook, and I was disgusted by the way the OS behaves now. There's NO way to get to the hard drive: they just hide everything and give you Documents, Pictures, etc.... The OS itself has removed hot keys and functionality in favor of "ease of use" but really, it's just being dumbed down.
The Mac OS was always for casuals, but when they added Unix, it changed and both casual and hardcore could love the same OS. But clearly the hardcore are no longer of any interest to Apple. It's a real shame because the Mac OS was very stable and useable for many years. Now, it's just a toy.
Oh, and Apple changed it so you need an Appstore account to update software. Real pain in the dick.
And finally, Apple doesn't need ZFS, necessarily. They need a standardized FS though. HFS+ is ghetto, and a drive formatted HFS+ is essentially useless to any other type of machine.
> The guy says he's going to list bugs, then he really doesn't, he just bitches in bulletpoint form.
Did you miss the pages following the first? He provides numerous specific examples. He even says in the original article (though it's easy to miss, which makes me think you missed it):
"The pages that follow are those personally encountered— not 3rd party reports."
I completely agree with some of what you're saying but it's criticisms like "There's NO way to get to the hard drive: they just hide everything and give you Documents, Pictures, etc...." that seem misguided: of course you should hide the inner workings of the OS from the average user, but it's absolutely trivial for anybody to see their hard drive if they want to (including hidden files). The easiest of which doesn't require any advanced knowledge, but is just a matter of going into Finder preferences and checking "Show these items on the desktop: [] Hard disks"
Criticising Apple for hiding things that the average user shouldn't know about seems to be getting in the way of completely valid criticisms like what you say about OS X needing a more modern FS.
Its somewhat strange to on the one hand argue that the OS should be super simple (so simple as to hide the hard drive from the user), and in the next breath describe changing a preference as trivial: its not, stuff like that is really confusing to precisely those non-computer people -- especially Finder preferences which is often confused with System Preferences or Finder's other preferences (known as "View Options").
The argument is not that Apple should not make things easy. It is that Apple is increasingly doing this in the laziest way possible as opposed to actually thinking about the problem. The hard drive is confusing? Oh just hide it by default and add a preference. Ugh. They used to hate preferences, now there's one for strange trivial stuff.
I realize this is flogging a dead horse, but the point is that the OS is super simple for novices/casual users, but is trivially adjustable for power users. It doesn't matter if those adjustments are confusing to non-computer people - that can be a good thing - but it's nice that with a bit of computer knowledge, you can essentially be running a solid Unix machine with a pretty, though not universally perfect, interface.
My point is that I don't think (from my own experience), that these things have necessarily been made any more understandable in OS X land by half-porting some of these iOS metaphors. My concern is precisely that the experience really has not gotten much easier for novices, but more convoluted by competing ideas of the way computers should work shoved together.
Let's look at the hard drive example in more detail. What did we lose by hiding the hard drive? Arguably the most important thing was access to Apps. This has been a critique of OS X for a while, that the main way to access Apps is this strange folder that lives in the hard drive. Perfectly fair thing to want to fix. Spotlight made it simpler to launch for pros, but you can't expect the primary way to launch an app to go from recognizing an icon to remembering the name and typing it into a textfield that appears when you click on the top right, especially if the goal is to move to a world where people buy LOTS of apps. So to remove the hard drive, you need another easy "visual" way of moving through your apps. What they chose to do was introduce the "application picker mode" to replicate the home screen of iOS.
But the problem is that this is NOT the home screen of OS X, and so the experience is jarring and scary vs. a safe place to return to. Perhaps had they done something bold and said, look, by default now this is what you look at, then maybe it would have been interesting. Instead it is literally a bolted on feature. In iOS, there is no question how to launch an app, you see them all in front of you. In OS X, you start with a strangely empty desktop and a few "blessed" apps in the dock, and you are supposed to divine that a chrome spaceship means "show me the rest of my apps" Or perhaps you are supposed to figure out that a five finger gesture does that.
I have not done extensive user studies of course (but then again, neither has Apple), but I have witnessed this feeling scary to people like my mother. They don't know how to get out of it and are afraid to touch anything. It is not the safe "home" of iOS, it is the exact opposite: bizarro desktop-is-blurred-out mode. Touching most things on the screen now leads to launching some application, and since OS X comes standard with a ton of apps you'll never use, the app they launch is probably a random app they've never heard of which results in me getting a phone call asking whether they should be scared because they launched Time Machine or Game Center or some other thing.
Again, my point is not that Apple is not worrying about novice users, its that these features feel exactly like a pro user lazily saying "fine they want iOS, here's iOS". Apple understood that desktop metaphors wouldn't necessarily work on a phone, that's why they took the time to think up new ones. It would be nice if they'd also realize that phone metaphors don't always work on the desktop. Or at least think these things through a little more before just dumping them in.
I 100% think there is a system that is 1000 times easier to comprehend for people and hides the hard drive in a way where the common user will never need or care for it. OS X is NOT that system.
Now that I think about it, perhaps the "the big empty desktop" should go away. We have a proper documents folder, so it shouldn't be a dumping ground for all kinds of junk, which is usually what happens.
I consider myself a power user, and I've trained myself to stop dumping shit onto the desktop. And now it's empty. After reading your post, I think it's a great idea for the desktop to be replaced by launchpad. Currently, I keep my applications folder in my dock (in non-stack form).
If they made Launchpad the desktop, and reduced the mouse travel between app icons (I don't know why there's so much space between them... Macs don't have touch screens), I'd be very pleased.
The file system hierarchy exists for a reason: to organize information. Having the UI hide that organization in the name of "simplicity" and display it as a flattened list defies explanation.
The users that can't handle a hierarchical structure probably just puts all files on the Desktop anyway, so what's the advantage here?
I agree, Apple can hide things to make the OS less complex. But what irritates me is that, as a user, I upgrade to the new OS, and the most basic of functionalities is hidden from me. They've literally taken something away. You cannot put a link to your compputer or hard drive on the desktop anymore. Is it not unreasonable to think that your users might want to be able to do this as you've had this hard-drive-on-desktop paradigm for well over 30 years now? I'm glad there's a check box to show the computer on the desktop now. I would never have found that. I mean, 30 years of interface paradigms teaching me to make an alias or simlink there was obviously just some mental retardation on my part...
> My wife just got a brand new MacBook, and I was disgusted by the way the OS behaves now. There's NO way to get to the hard drive: they just hide everything and give you Documents, Pictures, etc.... The OS itself has removed hot keys and functionality in favor of "ease of use" but really, it's just being dumbed down.
What the heck are you talking about? The only thing I can think of is the Finder defaults to not showing hard drives on the desktop, when it used to default to showing them. That's it. I can still go to Finder preferences and re-enable "Hard disks". I can also put them back into the sidebar if they're missing there. And of course the Go→Computer menuitem (shortcut: ⌘⇧C) still works as it always has. As does Go→Go To Folder… (shortcut: ⌘⇧G).
Thank you for the dose of logic. Truth be told, I hate not seeing it, but having turned it on, I use it less than I'd expected. If I could just disable the option if saving to desktop, home and work would be beautiful.
Sort of related to HD options, the default of having no display of remaining storage space is irritating to me, especially on the MacBook Air.
And finally, Apple doesn't need ZFS, necessarily. They need a standardized FS though. HFS+ is ghetto, and a drive formatted HFS+ is essentially useless to any other type of machine.
I could make the argument about Reiser or ExtN outside of *nix, or NTFS outside of windows. There's a better attack here to be made than this one.
See, the reason I want HFS+ gone is that NOTHING supports it. NTFS is supported on Mac OS, albeit read-only. NTFS is readable on most Linux's. EXT is readable by a number of OS's. And I don't want ReiserFS on my Mac. I don't want to be killed and stuffed in a suitcase by my filesystem.
I know aliases and such still work, but you cannot, as far as I can tell, make an alias of yer HD on the desktop. Clearly yer not supposed to, but it just irked me.
Thanks for the HFS links. I've never seen either of those before. Coulda saved me a lot of time and money!
HDDs with an SSD cache, but rather SSD/HDD hybrids that intelligently move data from one to the other)
Sounds like a cache. Care to elaborate?
yes, we're sadly not using ZFS but that's in all likelihood due to licensing restrictions, not because Apple thinks HFS+ doesn't need replacing.
With the amount of cash on hand that they have, this does not fly.
Can we stop pretending that Apple's tech is as good as their design and marketing? It's okay to have average tech in a nice, well-made package--the rest of the developers aren't going to make fun of you, don't worry.
It's tiered storage, previous only used in extremely expensive enterprise storage infrastructure. The canonical location for frequently used blocks of data is the SSD, and no data is duplicated on both drives (except while copying of course).
I have one in my new Mac Mini and I'm very happy with it. Combined with Time Machine backups it makes for a very efficient, very reliable consumer storage setup.
Tiered storage was available if you wanted it, it just has its own issues that don't necessarily make it better.
It's not going to be faster than caching[1] and suffers from a lot of potential data loss scenarios.
Yes, you get your 32 or 64 gig of SSD back, but who cares?
Tiered storage is used in enterprise environments to maximize the use of expensive fast storage, or insanely cheap very slow storage. Like say, really really high end PCI express based cards, or really cheap but high latency tape drives.
However, this makes a lot less sense in a consumer environment where consumer SSD prices are dropping 50% year over year.
They also make a lot less sense with only two tiers anyway.
So, to review,
1. Unlike the original claim, "fusion drives" are not new technology, or pioneering (since, as you point out, it's been available in enterprise storage systems for quite a while)
2. It's not clear it's actually better than the caching system others use.
3. It's not clear it makes financial sense.
[1] On the read side, you could use the same algorithm for deciding to move to cache as oyu do in tiered storage, and get identical performance.
On the write side, you can use the SSD as a data journal and perform writes as fast as you can write to SSD. In a tiered storage model, all storage usually starts out on the "normal" tier, so writes will be slower, reads may be faster.
If they use some weirder model (like data journaling to ssd), then it becomes equivalent to cache (in both time and space usage, not just time) , not "better".
I'm convinced that Apple wants you to think that SSD prices are still where they were in 2010-2011, both to sell you the "Fusion Drive" baloney and to make it seem like paying $900 for an UPGRADE to a 512GB SSD[1] is in any way reasonable.
Really? I could have got tiered storage on my new sub-$1000 computer I got last month? Which vendor could I get that from, Dell, Hp, Acer?
>It's not clear it's actually better than the caching system others use....
It's just a different technology with different characteristics, it's not a question of better or worse.
One obvious benefit is that you get the sum of the two drive's capacity. Also the two drives are on two different drive connections, not a single connection to a combined device, so there may be bandwidth efficiencies as well. But you just sit there stewing in your own distain. God forbid that anyone try anything different or unusual, that's got to be bad, right?
1. I think you are projecting a bit here. You come off as slightly crazed.
2. I never claimed it was cheap, i objected to acting like this technology is new. It's quite widely used in the enterprise world.
3. The cache model has the drives on two different connections as well. See Intel's Rapid Storage Technology.
4. I think i refuted the idea that the sum of the two drives capacity is useful in the current market (it's slower and not cheaper enough).
5. I'm all for different or unusual. I object to them choosing a marketing buzzword and pretending they've come up with some heretofore unthought of technology. I think I made that clear when I objected to calling it pioneering or new technology.
> Can we stop pretending that Apple's tech is as good as their design and marketing? It's okay to have average tech in a nice, well-made package--the rest of the developers aren't going to make fun of you, don't worry.
yes, we're sadly not using ZFS but that's in all likelihood due to licensing restrictions
What lead to that conclusion? You are aware that their are several commercial products available or that have been available that were based on ZFS, right? The CDDL license includes a patent grant.
IIRC Apple dropped ZFS around the time that NetApp was suing over it. Sun couldn't grant patents they don't own and indemnification probably wasn't financially feasible.
ZFS scrubs can detect silent bit rot and recover any data affected. ZFS also has very easy, filesystem-level snapshots. Filesystem-level deduplication might also be a benefit in the future, though is likely too memory intensive currently for a consumer device.
Basically, on OS X, it would allow Time Machine (including Time Machine local backups) to be implemented in a much more elegant way, though I think it's fair to say Apple's actually done a pretty good job of it with their current implementation.
Wow, that list of problematic Apple products doesn't even include the buggiest, unrecoverable-crashiest, worst-interface-with-a-bullet Apple product I have ever had the misfortune to be forced to use: Final Cut Pro X. With near-zero documentation and essentially required additional-cost "options" like Compressor. The reviews and feedback on this thing are absolutely dismal.
Of course, you could use iMovie. Which has a posterization bug so big you can drive a truck though it, making it of little value for HD. Still unfixed after at least three years. http://www.aibal.com/imovie-banding-posterization/
The biggest bug for me in OS X is the lack of emphasis on backward compatibility. I can pick up a copy of Windows 8 today and it will still run SIMCITY.EXE from 1992. A month ago I had to upgrade my entire operating system (and lose a ton of customizations I had made) because the new XCode wouldn't run on my two year old Macbook.
Agree++. I used Macs for a few years (SE/30, PB180) but software I paid good money for kept not being usable, and my favorite piece of software of all time (HyperCard) was dropped from the O/S. Forcing my software to become obsolete is the primary reason I stopped using Macs.
Many in this sub-thread point out developers were warned. I'm thinking as a user:
I continue to use Quicken 2000 that I bought in late 1999 on a Windows 7 system I bought last year. I know with MS that I can buy a piece of software and will be able to use it a decade or two from now if I want, without having to pay for upgrades I don't need.
I have a box full of Mac software that cost me hundreds of dollars. Completely useless and without value. I have no such box of useless Windows software (though to be fair, I've bought less after having been burned so badly buying Mac software 12-18 years ago).
You're definitely not going to be able to run a MacOS executable from 1992. There is less backwards compatibility in OS X compared to Windows, but to claim there is a lack of any emphasis on it is to forget the ability to run Classic in OS X, the transitional Carbon API for porting old apps to the new platform, the Rosetta runtime that ran PowerPC apps on Intel processors, and so on.
Granted, Apple isn't interested in maintaining 20 years of compatibility, so these features are no longer supported. However, they served their purpose at the time of their usefulness.
Xcode might be considered unique in that it targets Apple's latest platforms for development, so it's not necessarily surprising that it doesn't support an older version of OS X. I assume you were running Snow Leopard, which came out nearly four years ago.
I'm not sure why you bring up Classic, Carbon, or PowerPC. Rosetta was dropped from Lion (two releases ago) and support for Classic was dropped when Rosetta was added. MS Office 2008 (I think it was) stopped working with OS X Lion (2011) because the installer executable contained PPC code. Not very many people need or expect 30 years of backwards compatibility. They just want more than 3.
> I'm not sure why you bring up Classic, Carbon, or PowerPC.
Because it's not true that Apple doesn't emphasize backwards compatibility. I also pointed out that once those technologies outlived their usefulness, they were no longer supported, and that there is indeed less backwards compatibility than Windows.
By the way, Office 2008 ran on Intel Macs. You're thinking of Office 2004, which was PowerPC only and ran for 7 years before OS X Lion dropped Rosetta.
Macs transitioned to x86 in 2005, with Apple repeatedly warning developers to GTFO of any PPC code they have ASAP. MS Office 2008 released years later still containing PPC binaries, and support from the OS was finally yanked in 2011, six years after the switchover.
This seems pretty reasonable. What is unreasonable is that MSFT continued shipping deprecated code as the newest and greatest, long after any reasonable update cycle.
To be fair to Microsoft, they weren't the only ones. App developers were warned for years since OSX's first release that Carbon was a transitional API that would be deprecated in favor of Cocoa. This was in 2001. Adobe steadfastly refused to update, and in 2007 Apple stripped 64-bit support from Carbon-based apps as a beating stick to get major developers like Adobe moving. It worked.
From everything I've seen developing both OSX and iOS apps, Apple gives plenty of notice between deprecating functionality and actual removal of binary support (from the above examples, six years).
Compare with my days developing Windows software, where developers knew well that Microsoft had zero teeth behind API deprecations, and will shamelessly keep using it for eternity, making future backwards compatibility efforts ever more painful - you will find code written in 2008 that uses functionality deprecated in the late 90s!
He didn't say that Office 2008 didn't run on Intel -- he said that it ran on PPC, which it did. 2004 was PPC-only, 2008 was a universal binary, and 2011 is Intel-only
That very well may have been a factor. However, Apple engineers have publicly stated that for them to continue supporting PowerPC Macs would have required developing and shipping system APIs targeting a processor they hadn't sold in over half a decade. I think Rosetta's number was up regardless.
That's forward compatibility, not backwards compatibility. Will the SimCity being released next month run on Windows 3.1? (Of course not, that's an intentionally extreme example, but you're fundamentally talking about two different things -- new app on old OS vs. old app on new OS).
You're right, bad example. A better example is that I spent a decade on VC++ 6.0 without upgrading and could still build software and use SDK's for and from future versions of Windows.
On OS X it started with wanting to test an iOS 6 feature, which required the new simulator, which required the new SDK, which required the new XCode, and before I knew it I was upgrading the entire operating system.
I'm not sure if you're aware of it, but Apple isn't on the PowerPC or MC68k arch anymore. Would you expect them to continue supporting apps that were made to run on the PowerPC & Motorola 68k chips too?
> there is no technical reason why the latest iPhone SDK cannot run on Snow Leopard.
How exactly you know that?! Maybe new SDKs (that are compiled to run on x86 for iOS Simulator) use ARC exclusively which doesn't work properly on Snow Leopard. It makes sense to do so now that ARC's been out for a couple years...
What application released two years ago isn't running on Mountain Lion today? The only one I could think of would be Xcode.
> there is no technical reason why the latest iPhone SDK cannot run on Snow Leopard[0].
Snow Leopard was released in 2009. Just because something can be hacked to run on an unsupported platform doesn't mean you won't encounter an issue. It's not just about dropping a platform as a target; it's also the fact that it won't be officially supported (meaning no guarantees or bug fixes), Apple's engineers won't be going out of their way to make sure the SDK and the simulator run, and they won't help you with technical support. If it does work, and there aren't any known issues (which I doubt), it's a happy accident that could break with any update.
My Quicksilver app hasn't changed since 2008. It's an old app but it still works being 5 years old. Maybe you're running poorly built apps and you should be complaining about shoddy developers instead of a shoddy operating system.
The execution is spastic but the intent is correct: substantial portions of the Apple codebase are lipsticked pigs.
Mail.app is, to my mind, the single most flagrant troublemaker in small-biz use. But there are thousands of little cockroaches found all over the place and I can't stop seeing them anymore.
here we go!
-- Doesn't (officially) support HTML signatures.
-- Requires a lengthy reindex on any internal database change, which can happen multiple times a year.
-- Severe, egregious issues with corrupted or blanked out Keychain entries
-- Easily works itself into a lather where it cannot properly authenticate with anything until the entire system is restarted
-- GMail IMAP sync-thrashing and message corruption that can render some accounts unusable
-- Crashes on queueing massive message previews
-- Customers almost uniformly loathe the recent UI changes
-- The new search UI is nearly unusable for most, requires frequent spotlight rebuilds
-- Drafts-on-server duping
-- Plugin hassles and failures, especially as regards Archive Utility and PDF handling
-- Evaporating local folders
-- New rules about Archiving that don't play nicely with grandfathered Archives
-- Auto-setup of new accounts that gets hung so badly you have to back in a patently entry just to get the chance to hand-edit settings
It is unreliable and unfit for heavy duty use. I've lost a third of my customers to Outlook.
Yes, I don't know if it is the Finder itself, or so other part of the OS/kernel that is in fault, and NO, I don't know how to reproduce it because it is such a scary experience (happend two or three times if I remember correctly) that I don't even want to go explore it, since I never even received an automated acknowledgment email from apple for bug reports or dump posts or anything like that. One time, there was a directory deep inside the one that was beeng moved that had the same name as the top level (I don't know if this was related to the error though) and midway there was a useless error saying some error happened and half the files could not be found not in the source not destination.
> This problem is only an issue for those doing certain types of upgrades; it won’t occur with as-shipped Fusion setups. But it shows an inattention to detail with severe consequences— many users do modify their systems.
So if you go way out of your way to DIY a Fusion Drive and set up all of the pieces required, it becomes Apple's fault when Disk Utility fails to acknowledge your heavy modifications?
I wonder if the author gets mad that you can't move files out of /System and still expect the Mac to boot. After all, many users do move files.
Even with a DIY Fusion setup, Disk Utility unhelpfully offering to wipe all your drives to fix a filesystem error points to a pretty major bug in Disk Utility's decision-making capabilities.
I think part of the problem w/ performance issues is that Apple doesn't seem to give a crap about older hardware. The upgrade from snow leopard to lion pretty much left my mid 2009 macbook pro unusable. After many miserable months I gave up and switched over to a linux desktop for full time development work.
However, my brand new shiny 2012 macbook air _flies_. It's about as fast (build times, and what not), as my 2009 era desktop running linux. I have had similar experiences with my iphone hardware. You have to keep upgrading hardware if you want to keep up with the latest software.
Given that, there are still some awful bugs in later incarnations of OS X that remain unfixed. For example, lion's retina support broke external 30" displays in landscape mode. This is still not fixed 2 years later, despite bug reports and a very easy to reproduce problem. Apple just doesn't give a f*ck about power users anymore.
> I think part of the problem w/ performance issues is that Apple doesn't seem to give a crap about older hardware. The upgrade from snow leopard to lion pretty much left my mid 2009 macbook pro unusable.
Just curious, in what way did the upgrade make you MacBook Pro unusable? I had a 2007 MBP that used up until August 2012 running Mountain Lion and never had performance issues. Eventually the keyboard died and rather than repair it, I sold the laptop and replaced it with an 11" MacBook Air (which has been great thus far).
This is a very subjective and vague list of criticisms. It reminds me of the iPod haters around 2003 who didn't like that Apple was spending so much time on its music player.
I mean, what exactly is he talking about when he says OS X is becoming more of an entertainment platform or that features are getting "dumbed down"? No examples are given, so apparently it's just a general feeling he has.
Due to its lack of details, the article isn't persuasive.
I personally have been feeling the quality bar slipping but this article goes way too far. For me things aren't that bad but I can see where it's heading. Hoping to get another 5-10 years out of the Apple ecosystem.
Hardware support should be better. I've had at least four firmware updates since the Air was released, which seems extraordinarily high for a product in a closed ecosystem.
And by the time that 5-10 years is up we will probably be in the middle of a huge paradigm shift. I am not sure of the details but I think everyone can sort of sense it on the distant horizon.
Yeah, OSX is subject to consumerisation. So is Windows. I don't think it will get too bad though - after all, the Apple developers have to use it. I think we just have to kind of roll with it. The consumerisation of computing has a momentum all its own and I don't think anyone knows where it is going to end up.
>>Having been a professional software engineer (C++, Java, assembly code, drivers, compression, encryption, threading, server code) for 25 years, and having used Macs since they first appeared back in the early 1980’s, I have a long and deep perspective on the evolution of Mac OS X (now just “OS X”).
I expect A LOT more then that is that is really what he did.
Probably talking about the old carbon threading apis, which were depracated along with the rest of carbon in 10.8. I'd be surprised if anyone has been using them since 2006.
This is a more general trend in computing. iOS and similar are much too popular with "normal" people for anyone to still care about having a real general purpose machine.
This article is dribble. Some of these are issues some are not. All of this is hyperbole and has little to with reality. Phrases like "can't copy" are a red herring, as copying is easy and reliable, HFS vs. ZFS, Etc etc. Very few, if any facts or evidence.
Yeah, it's a little bit of a 'turd in the punchbowl' with regards to OS X criticism. I stand by my opinion that Snow Leopard was the last version of OS X that I felt truly enhanced my experience, but a lot of what he says in the article is nonsense.
Snow Leopard was indeed the high water mark for OSX, and probably will be. It will be a real shame when SL is end of lifed and 3rd parties no longer target it, etc.
> Outright removal of an API in a minor release. Deprecation with threat of removal of robust long-standing threading APIs with rewrite required.
Hasn't Apple always done this even before OSX? I could be wrong, but I vaguely remember Joel Spolsky writing about this as a core difference between MS and Apple.
He says the platform is going downhill and becoming entertainment only, then mentions a bunch of stuff that is either a new addition or media-related. E.g. iTunes doesn't suck any more now than it ever has, nor does the Finder.
Relative to OS X 10.2, 10.8 has:
1) full screen support, which is a huge advantage to coding on a constrained laptop screen; the newest Emacs supports it, so you can easily swipe between your editor, your browser, etc.
2) much better and more complete POSIX support;
3) an entirely new tool-chain in Clang/LLVM; XCode sucks, but it has always sucked;
4) fleshing out of little low-level features (improved Obj-C runtime, thread local storage, etc)
I just don't see anything non-trivial that OS X 10.8 can't do that say 10.2 could do, and a lot that's the opposite.
"OS X is degrading into a base for an entertainment platform. As it stands, the trend is entirely downhill for serious work..."
->Interesting. I see more and more developers switching to Mac and taking it seriously. My last office switched completely from Ubuntu to Mac.
"OS updates are fast and furious— a lot of hype but little of real value..."
->The real value is stability. I don't care for a 100% revamped UI every 4 years like Microsoft pushes. I like a familiar platform that is steadily improved.
"Core operating system quality is declining as resources are diverted to software development in more profitable lines"
->The core OS is Unix. Unix has had decades to get it right and I doubt Apple is somehow ruining it in the last few years. Yes, Apple focuses attention to the consumer appliances, but I feel that's mostly because investors and press are obsessed with that so that's just what it appears to us. I suspect they fund Mac OS pretty well.
"We begin to tread in dangerous territory: potential data loss"
->I agree here 100%. I've seen too many Time Machine problems.
"Developers are forced without recourse (by API changes and Apple Store requirements) into costly and arbitrary updates which themselves carry the risk of new bugs."
->I wouldn't know. The great thing about Mac is that it's a good OS for webapp development which I focus on. So I encounter none of these issues.
"Useful functionality is prohibited in the name of security."
->Have you been around in that last 20 years? Do you like ActiveX and Applet plugins in browsers too?
"Outright removal of an API in a minor release."
->Well, if you know that, isn't that half the battle? Pay attention to minor releases then. Maybe since they can't really upgrade beyond the major version of 10...
"Apple’s iron hand over what constitutes a 'right and proper' application leaves no room for disagreement."
->Agreed this is a step backwards. But the blame is partly on users for accepting this.
"Hardware for professional use is released in 3-6 year cycles (Mac Pro), or is dropped entirely (XServe and related)."
->All the professionals I know like the MacBook Airs increasingly. Hardware has stagnated overall in the industry. Apple has pushed forward with thinner cases, NAND drives, and Retina displays: much more innovation than competitors.
"The trend to a new breed of 'shallow' features: those useful only for beginners and entertainment"
->Bells and whistles sell. So sue them. They got the core, essential features done on the first OSX release.
"The general dumbing-down of features in every Apple OS X program."
->I like this. The default OS X software is more about managing your life. I don't need a fancy shortcut to sync my camera to iPhoto (no idea if this exists). I need shortcuts for development where it actually saves time since I rebuild 50 times a day, and Xcode for instance has configurable shortcuts.
"The general trend to introducing stupidly inappropriate iOS-isms into OS X (e.g., Mission Control)."
->Really?
"The OS X donkey cart is getting loaded with ribbons and flyers and decorations and marching band"
->Did you run out of ideas for the article at this point?
"So-called OS X 'upgrades' now consist largely of ill-conceived dilettante eye-candy features that reduce usability, clutter the user interface and introduce scads of new bugs."
->Apple QA could improve. But I think their UI's have steadily improved which is important. Look at the first release of OSX to now: much better.
"The real talent at Apple has probably been diverted away from OS X to iP* development"
Not in my experience. Air keyboards are horrible. The older (pre 2009) keyboards were way, way better. It's easy to accidentally press a nearby key if typing fast, so I have to type more deliberately (and slower). Maybe I just have fat fingers. Thinkpad keyboards get it right (for a laptop). I prefer everything else about apple hardware, though.
I don't see how the ergonomics of the Air would be significantly different than other laptops. It has slightly less pixels on screen, but the screen height and keyboard size remains similar to all other laptops. The only way to fix ergonomics on any laptop would be to raise the screen to eye level and use an external keyboard at arm level.
Hey! Heard about external monitor, keyboard and mouse? I use 2011 MacBook Air, has actually used a MacBook Air since the first one. Had a Mac Pro for some time, but now I only use a MacBook Air. Depends on what you do, but works grate for me and I am a "power" user, linux VM and Windows VM running at the same time, MAMP, Photoshop etc.
I don't understand this argument. If you use external monitor, keyboard and mouse, then the fact that it is a macbook air or a dell with os x pried in it no longer matters.
There are degrees. ThinkPad keyboards are fantastic, the Air is a mess. I have a Vaio Z which a keyboard even worse.
It's true, though, that for serious long term use you really don't want to be using a laptop at all. My Vaio goes on a stand at work and I use an external keyboard (Microsoft Arc, which I was surprised to like very much). Hunching over a laptop to type is just a terrible state to be in.
The key sizes on the Air are identical to the Thinkpad, so I don't see how the ergonomics would be different. Your hand position on the Air would be identical to the Thinkpad. I don't think there are any laptops out there that would have an improved keyboard for the purposes of ergonomics.
You don't think the key action matters? Try typing on some different keyboards. The Air (and my Vaio) have very short-action compression strokes and very little tactile or aural feedback. You have to press harder to get them to register than for a longer-stroke keyboard, and yet your brain gets less feedback that you've succeeded.
Basically, spend some long days typing on a bunch of different laptops. Obviously everyone is different and taste matters, but unless you're a 2-finger hunt-and-pecker, I all but guarantee that you'll find you like some keyboards much more than others.
And I've found that the low-profile Apple keyboards are the very best. I simply cannot stand long-action keyboards. An Apple keyboard feels like I'm walking on water. Long-action keyboards feel like I'm wading through mud.
I don't think key action matters much to ergonomics. It doesn't change the positions of your hands or fingers. It's like comparing the health benefits of orange juice more pulp vs less pulp.
I think this is pretty subjective and personally I disagree with it.
I have a both a MacBook Pro for work and a personal 13" MacBook Air. I use both connected to external monitors but with the internal keyboard and track pads. The keyboards seem pretty similar to me.
I touch type with a non-qwerty layout (so no point hunt-pecking) and I find I really like the short, punchy travel of the keyboard. For comparison I also have an IBM M keyboard that I no longer use.
I guess if you want long travel keys they're not for you but it's, like, just your opinion, man.
When I was young OSX, was so much better. It used to read my mind and produce everything I needed even before I'd thought of it. Including coffee, and sweet blueberry muffins. Without even plugging it in or turning it on. It was so much better than Windows.
These days, it is merely an operating system.
I may switch to Arch Linux, which I hear has a voice-activated Cappuccino machine and dry cleans shirts.
> OS X is degrading into a base for an entertainment platform. As it stands, the trend is entirely downhill for serious work (albeit a mild grade so far, but steadily downhill nonetheless).
When was OS X not an entertainment platform? iPhoto, iMovie, iTunes -- these applications have been around for a long time.
OS X is less of an entertainment platform than Windows ... when you look at the fact that games are still almost always released for Windows only/first.
OS X was and still is a platform for creating, iPhoto, iMovie, iDVD, iWeb, and I would even go so far as putting iWork into that same category. Windows has Windows Movie Maker but it doesn't compare to iMovie in terms of output and the kinds of things the average user can accomplish with it.
I think you are both missing the point. The author is trying to argue that Apple are removing features and "dumbing down" the OS, to make it easier for casual users to use it as an entertainment hub.
There's also a lot more to entertainment than games. iTunes is the dominant media application, and it's so much better on OSX than Windows. A lot of people consume their entertainment online nowadays anyway, my kids are growing up that way.
Between Steam for Mac and the Mac App store availability of top tier games on the Mac is excellent and getting better all the time. Can't play Far Cry 3? Oh well, I'll just play Portal 2, CoD Black Ops or The Darkness II instead and there are scads more.
+1 for your point about creating though. Both my kids are very musical and are playing with Garageband on the iPad already, and when they're a little older it's there for them on the desktop too. My eldest is getting into photography as well and while she may well outgrow iPhoto at some point, it's hard to imagine a better way to get started.
But then I'm squarely in Apple's largest target demographic for the Mac. That doesn't mean the mac is only for home users though, any more than Windows is only for gamers and Enterprise users. It's just that those are their particular strengths.
Don't the bullet points all repeat themselves. It could have been just:
* Software products are being dumbed down for novices at the expense of power users
* OSX isn't getting enough attention because the iPhone and iPad are so big now
Things like kqueue() support not being up to par with the other kqueue() providers, notably FreeBSD. kqueue() is considered so broken by libev for example that it just doesn't use it. The deprecation of OpenSSL (although I'm happy about that, but installing a secondary OpenSSL alongside the OS provided one can break things, so you have to install to a separate path and use include directives in your compile line, and or an alternate pkg-config path). The fact that a replacement libc++ now exists, but parts are still using the old stdlibc++ and the two can't be linked together if you use std::string for example, thus making it harder to use C++11 features. This is going to be an issue for a little bit, but hopefully that gets better. Outdated utilities or libraries make life more difficult, however a lot of that can be fixed with homebrew. Debugging USB issues is more difficult than it should be, this is especially an issue when attempting to figure out why a driver from a vendor crashes when you use certain USB to serial devices.
The new sandboxing requirement for apps while making everything a little more secure makes it more difficult to make certain types of apps ... slowly but surely new API's or permissions are being made available to fix those issues, or at least improve the pain points.