Apple's marketing is an added feature. I get a lot of joy looking through their ad copy, in a way that I get out of very few marketing departments'.
I always wonder why Microsoft is so bad at advertising. They're not a stupid company. They create products that are nearly always good, if rarely (in my opinion) great. But they've never shown any sort of strength-of-focus and they make things that are ugly. I don't get it. In cases like this, it's easier to make something that's stark and simple than it is to make something cluttered and unappealing.
The difference you're seeing is design that's driven by an individual with a single strong voice versus a committee. For clarification, I'm saying that Steve Jobs designed this web page, but whoever did was someone with the authority and the balls to say, "no" over and over to many different people. Those Windows 7 pages reek of too many cooks in the kitchen and/or lack of leadership.
As a non Mac OS user, that link just reinforces to me that it looks like just a service pack. But at $29, I don't think it matters. If it were $129, that would be ridiculous.
As a user, yes, but as a developer, there's some really awesome stuff in this release (both from the perspectives of the apps being developed and from the tools used to build them). Can't wait.
I can't believe it has been considered important enough to be one of the major features of Snow Leopard, I am so happy for the blind and visually impaired. They did it with the new iPhone the first commercialized accessible touchscreen and they are doing it again with the new Mac OS. Apple is the most innovative tech company there is no doubt about it.
True, but unlike practically everyone else they didn't allow their work to fit the expanded time available. Everyone builds in buffers and pads their estimates, but few people are able to work to their original true estimate.
I wish other companies would follow OpenBSD, in regard to releases, package management, etc. Their community is awesome. Abrasive at times... but awesome ;)
True, but unlike practically everyone else they didn't allow their work to fit the expanded time available.
Again, how do we know they didn't account for it when they buffered initially? Maybe they had a different internal timeline (end of aug) vs. an external timeline.
Only way to know is if one is in the Snow Leopard team or has access to that kind of info.
> how do we know they didn't account for it when they buffered initially? Maybe they had a different internal timeline (end of aug) vs. an external timeline.
If they were like practically everyone else in this wacky industry, they would be late for the internal timeline but not worry about it because the external timeline is the one that matters. Thus, there would be very little pressure to meet the internal timeline and a lot of pressure to meet the external timeline. Worse, they would wind up being late for the external timeline as well as the internal timeline.
For example, if someone wanted to add a feature to Snow Leopard, the dev team might say "we can't make our internal schedule if we add this." But some product manager would say "well, will you still make the external schedule?" and bang they would start to slip.
If they had an internal schedule and managed to stick to it I believe they are an exception to how software is usually developed.
While I'm sure the scheduler had to be changed to accommodate it, it's not primarily a new scheduler. It's really a sophisticated thread pool system implemented in the kernel and exposed to programmers through language extensions.
Thanks - this makes more sense. Just a scheduler wouldn't explain how "With GCD, threads are handled by the operating system, not by individual applications. GCD-enabled programs can automatically distribute their work across all available cores, resulting in the best possible performance whether they’re running on a dual-core Mac mini, an 8-core Mac Pro, or anything in between."
Yeah, that sounds about right. It's a thread API that relies on applications providing "blocks" of work to the OS, and transferring their arguments and results around in message queues. So if you have five applications running, each of which has 8 threads to handle the needed parallelism on a Nehalem box, you don't need 40 separate threads and their stacks. The OS just spawns 8 threads.
That sounds OK, if kinda unexciting. I'd like to see some benchmarks vs. traditional threading before I commit to liking it. History is littered with cute new IPC mechanisms that didn't turn out to have the benefits promised. This sounds a lot like a combination of sysv message queues and solaris doors, neither of which managed to drive much real innovation.
The thread pool approach allows you to exploit much finer grain parallelism than explicitly fork/joining your own threads. So, an application that makes good use of this would probably have hundreds, if not thousands, of discrete work tasks (which I normally call "units of work," but Apple calls "blocks.")
What a thread pool infrastructure buys us is the separation of tasks (the bundle of information passed around the queues) from their execution context (in this case, a kernel thread). It puts a level of abstraction between what is executed and how that's executed.
At the risk of belaboring a point, this is an old idea - this is traditional multithreading. Java has a library for doing it, as does C++ (in Boost). A library for this kind of parallelism was, in fact, the first C++ library, written by Stroustrup. Even the Linux kernel uses this technique internally (by which I mean it's not exposed to applications at the user level). Cilk (the MIT project that spawned the company Cilk Arts which was recently acquired by Intel) is probably the best known example of providing language-level abstractions for this technique. (I reccomend their paper "The Implemention of the Cilk-5 Multithreaded Language": http://supertech.csail.mit.edu/papers/cilk5.pdf)
By pointing out that this is an old idea, I don't mean to denigrate what Apple has done. Execution matters, and from the looks of it, they've executed this well - they've got something working at the language and kernel level, which is non-trivial.
Actually (and I'll hide my lack of knowledge by being vague) it's a means of making it easier for developers to create robust multicore capable software. I think.
.NET has had runtime-wide thread pools since its inception. Unless there's something fundamental about GCD that I'm missing, that puts Windows at least 7 years ahead of OSX on this count.
It's interesting. For the little I have seen (and I was only able to give it a quick glance) it includes syntactic support for multi-threading being built into their Objective-C compiler.
Chris Lattner's first paragraph here talks about that a bit, although it doesn't give a timeframe for when the gcc implementation will make its way back to the main distribution:
Although, there's a comment by Johannes Fortmann here that says that the code has been available in an svn repository since at least September 5, 2008:
Sure. I'm not trying to say that Windows was ahead of its time, but rather that Apple seems to be making a big deal out of a feature I'm kind of shocked they didn't already have.
But it then adds in x86_64 support, which brings the size back up. The smaller files are likely to be from not installing all the resources (localizations, drivers, etc) at once if not needed.
May I ask how it is "now available"? From the site, it says that it will be delivered Friday. Did Apple not update the page this links to? Is this an "it's available for purchase now for delivery later" post that the headline is misleading me on?
Anybody know if you can install this on a clean hard drive? I've got Leopard on my MacBook, but want to swap out the existing drive and start fresh on a new SSD. Would be nice to not have to do a Leopard -> Snow Leopard dance.
Technically, it does (cheap upgrade copy of Tiger for those, who purchased Mac with Panther after the announcement of Tiger, required Panther). Still, I believe it's true that you can do a fresh install.
You might want to hold off a little while before upgrading, at least until 10.6.1. I upgraded my laptop 3 days ago and since then I've had at least 5 Mail.app crashes, finder a couple of times, and one terminal crash. I did an upgrade install, maybe a clean one is better, but I'm pretty surprised at the instability.
Also, if you use Macports, that's not up to speed yet so you'll be doing a lot of fiddling around to get that working. For example, I had to patch the port for Erlang to get that to install on 10.6, there's probably others. Wait a few weeks and that should all be sorted out for you.
I've been using it for a week on a 1st gen MBP (32 bit) and haven't had any crashes.
I don't use Erlang, but I have a lot of packages from MacPorts (mysql5, mercurial, git, svn, xorg, perl5 and python26 ) installed without any problems. The only thing I've tried to build that failed is screen.
I did a clean install, so that may be the difference maker.
Could your instability be attributed to the fact that you might not be running the actual release version (or has been modified in some nefarious way)?
I should have mentioned that I also had the md5 of the release image. Needless to say, I verified it. There is no practical possibility that it is not bit-for-bit identical to the retail version.
I checked the file in the same way you would ascertain a Linux iso you downloaded from an unknown source over the internet is genuine. There is no doubt whatsoever about its origin and authenticity.
I'm still on Tiger and I'm probably not going to upgrade. I have a first-gen "lapburner" MBP which works just fine. Leopard didn't really offer anything ground breaking and it doesn't appear Snow Leopard will either. And I figure anything ground-breaking will require newer hardware, so I guess I'll stay on Tiger until the machine stops working. Which coincidentally is also why XP users don't upgrade either.
Insanity. Before Leopard, Spotlight was a debacle. After Leopard, it just works. Also: isn't there a whole lot of software you can't install now because Tiger doesn't have Core Animation and isn't LP64?
I also wouldn't trust Tiger security on an external network.
Every security update Apple ships for Leopard it also ships for Tiger.
And I preferred spotlight in 10.4. In fact, 10.4 was all around a much better OS than 10.5, which was the first update to Mac OS X that felt like a step backwards.
Of course, I'll be updating to 10.6 as soon as I can get my hands on it. If nothing else, the promise of a 10.5 that doesn't suck is worth it. To be fair, 10.5 is much better now then when it shipped. When it shipped, it was terrible.
How was it a step backwards? Mind you, I was a bit late to the party - I started on 10.5.2 - but everything's much more polished than it was in Tiger, both graphically and performance-wise.
Leopard, pre 10.5.2, was indeed a mess. Lots of performance problems. I don't think it was until 10.5.5 that it caught up to Tiger in terms of performance and stability. 10.6.0 is great out of the box.
I think it depends. If you have a first-gen MBP than you are better off getting the new OS. 10.4 "supported" Intel chips but it always felt like a hack between PPC and Intel platforms. 10.5, and each update since, has gotten progressively better for Intel chips.
I think you've fallen for the marketing trap where a list of features has to be present for it to be worth something. Not many lists include code and platform optimizations because it doesn't market well. I've used OS X since 10.1 and each update has gotten remarkably better. Things have mellowed out in 10.5 but the OS feels fully baked.
Good points, but I don't think I'm falling for a marketing trap to want something tangible for paying out $130. The OS works fine for me now. If Apple wants more money for me then I want more than just optimized code.
In the end I don't think there really is anything they can do to make me pay more money for the OS. I already bought iWork 08, I bought SuperDuper! for backup, and everything else I use is either open source or comes included with the machine. If Apple wants me to upgrade, then they need to discontinue security updates for 10.4 or some widget I badly need needs to be 10.5+. And honestly, it's not me that Apple has to contend with, it's my wife and her iron grip over the budget ;)
Time machine alone was worth the price of admission. There were also some good unix upgrades in there if that sort of thing carries any weight with you.
Time machine is great, except for the fact that it Denial-of-Services my machine every time it does a back up. It reminds me of the late 1999-2000 when the Anti-Virus programs for windows (Norton/Symantec) would do a full system scan on a desktop and make it useless for an hour. That, and my 18 month old top-of-line spends an inordinate amount of time with the spinning beach ball. And crashes when I plug in a 30" monitor 20% of the time.
I'm hoping that a scorched-earth fresh install of 10.6 will either eliminate the amazing flakyness or alternatively point me to a hardware problem.
That's not exactly a good reason to upgrade by itself.
It could be if you were required by some regulation to use a UNIX certified OS. IIRC, there were little feature changes that warranted the certification.
As for myself, I have an older iMac that won't run SL, a netbook that runs Linux and I am perfectly happy with both.
The difference is that Mac OS X tends to get faster with newer versions. Apparently Snow Leopard even takes up less disk space. Not the case with newer versions of Windows.
I wonder how its going to work. If you have Leopard and your hard disk dies - do you have to install Leopard and then upgrade to Snow Leopard (which would be a pain), or will the upgrade allow a clean install onto a blank hard disk?
Still on Tiger here too, which is my main reason for wondering if I can get away with the $29 version!?
Reports indicate that the $29 version will install on a clean machine just fine, no need to install Leopard first. It's kind of an honor system thing with Apple.
Cheers for the info. I will get round to upgrading eventually then, but I reckon I will give it a few months until any compatibility issues are ironed out!
Hmm, not exactly convincing. Leopard brought stuff like time machine, this smells more like a big service pack. I mean, OpenCL and some of the central applications running 64-bit? Better Quicktime and a new Safari? I don't even use QT and Safari.
And it's $29? Too little to be too much, but still a lot for a service pack.
Apple choosed speed, optimization and refinement over features. And I welcome this choice wholeheartedly, enough bloat already... make my OS faster, slimmer and more robust and I'll be happy.
Sadly the price point of Snow Leopard, and the sort of reaction that you have make me think that the big shops will continue to make ever more massive and bloated softwares...
After all, you get what you are ready to pay for : if you think features are worth more bucks than optimization, that's what you'll get.
Here http://www.apple.com/macosx/technology/ are the new features that went into this release. And your comment doesn't even make sense, first you say it's a service pack than you rattle off a short list of new and upgraded items. These things do cost money to develop. And I'm pretty sure $29 barely covers the cost of it. I don't remember anything new being in Windows when a SP rolled out except a new browser. (There was the uPNP browser that popped up in SP2 I believe)
Every other "system update" reboot I've made for the past 20 months seems to have included an update to QuickTime and Safari -- both of which, by the way, are available as free downloads -- and the prominent featuring of those two products as a component of a whole new version of an OS makes it feel like a service pack.
> and the prominent featuring of those two products as a component of a whole new version of an OS makes it feel like a service pack.
point me to a page that only displays those 2 updates as the main features. In my browser Apple is clearly marketing the 64-bit, Grand Central Dispatch and OpenCL as the main reasons to upgrade.
Also stop referring to an OS upgrade as a service pack. Because it isn't. Microsoft releases service packs, just like Apple releases minor updates to their OS, (eg. 10.5.7)
Thras, I'll ask you this completely sincerely, because I don't and never have understood this. Why do you, and people like you, get off from reading a lengthy debate on a topic online, and ignoring every point except the one that's not a point, then snarking about it? It's not just you: A lot of people do it, and it baffles me. I always really enjoy being told why my opinion on something is wrong, and these snarky responses don't seem to prove anything.
You more or less said that the grandparent was wrong "just because." You were telling someone that he was wrong without any argument. Exactly what you complained about with me, btw.
The only possible response to that sort of thing is snark.
When Apple first announce 10.6, they said it wouldn't have any end-user features, but just a lot of re-engineering under the hood. Well, they couldn't help themselves, and there are a bunch of smaller features, plus things like Exchange support across the board.
And there really has been a lot of re-engineering, most of it focussed on making things a lot faster and smaller.
I think Apple deserve our thanks for actually making performance and improved footprint an explicit goal of a major release, rather than larding on new feature after new feature in some kind of misguided feature war.
It depends on what you plan to do with it, I suppose. I'm personally happy to pay merely for a decent chunk of performance improvements (isn't that what we do when we buy faster computers, after all?), but the new frameworks are also interesting—OpenCL, for example.
http://www.apple.com/macosx/
http://www.microsoft.com/windows/windows-7/
I mean seriously, how many levels of tabs are too many?
http://www.microsoft.com/windows/windows-7/features/whats-ne...