Slashdot Mirror
OS X Snow Leopard Details
JD-1027 writes in to kick off a discussion of OS X Snow Leopard. Apple's stated goal: "Taking a break from adding new features, Snow Leopard — scheduled to ship in about a year — builds on Leopard's enormous innovations by delivering a new generation of core software technologies that will streamline Mac OS X, enhance its performance, and set new standards for quality." The technologies: Grand Central to get better use of multiple processors and multicore chips, OpenCL to tap the power of the GPU, 64 bit so we can finally have our 16 TB of RAM, QuickTime X for optimized modern codec performance, and built in Exchange support in iCal, Address Book, and Apple Mail that most likely will help get Macs into corporate environments. We've previously discussed ZFS in the server version of Snow Leopard."
...if this will be a free upgrade similarly to the upgrade from 10.0 to 10.1. It would seem hard to justify a purchase price of anything more than $20 that adds only additional stability and developer tools. If anything, this version seems more geared for developers than end-users.
Make sure everyone's vote counts: Verified Voting
It is about time. We have zillions of programs for every major OS; so why waste time and money on adding features to the OS while third-party already do it? I believe it's a clever idea to enhance the core OS while keeping the outside intact (no new feature). Microsoft tried it with Vista, and they failed miserably. Was the task too big? Maybe. I hope Mac can achieve a complete OS core overhaul in a timely manner. It would set the bar pretty high for other OSes.
Jobs announces he's going to enormously simplify the morass of parallel programming and then also take GPU programming languages far beyond NVIDIA. And he's going to make this all in the core of the OS so it will be ubiquitous.
Oh and one more thing, we've already done it and it's going to be in our next release
Then I read posts about "well what about NTFS or Power PC".
Jebezus! get a sense of proportion here. Yeah NTFS might sell a few enterprise computers. So maybe that matter financially. But apple's doing fine with it's cash flow and we won't be talking about NTFS 5 years from now.
We will be talking about the future of computing which is how to tame and unify alternative and multicore architectures in a way the programmer does not need to worry about.
That's earthshaking if it could be done next year! Now a lot of people have blunted there spears chargin at this one so one needs a healthy dose of skepticism that it could be accomplished in a decade let alone in a few months. On the other hand the one person we know not to scoff at when he says he's going to make something complex really simple, retain 99% of it's power, and deliver it ubiquitously and accessibly is Jobs/Apple.
So doubt and wonder. Pour awe and skepticism. But fuck, don't ask about NTFS when this kind of thing is being annouced. You might as well ask about Zune support in Itunes.
Some drink at the fountain of knowledge. Others just gargle.
I don't see why they would drop PPC support yet. Certainly, stripping PPC code from an Intel Mac doesn't make much difference to the disc space use. Mostly, stripping out unused languages makes much more difference. I gained 2.5GB of space on my MacBook Pro by doing so and I now have universal binaries that are very similar in size to those seen in Snow.
They still have to maintain a port of Mac OS X just in case, and the also have to keep OS X running on the iPhone (Strong ARM) so I don't see the benefit of focussing just on Intel CPUs. In addition, keeping code running on PPC will help with keeping bugs down as it is often the case that just the act of compiling C code for a different architecture can result in unseen bugs showing up. As for performance tuning, rarely do you need to worry about much more than some small parts of the code to fine tune for a specific platform.
I'm not surprised that this developer preview is Intel only but I will be surprised to see the final release be Intel only. Leopard on PPC could no doubt do with some fine tuning although it does run surprisingly well on my nearly five year old G4 iBook. Besides which, the last of the PPC machines were being sold by Apple as late as the end of 2006 (PowerMac G5s) so I think it would be a bad move for them to drop support this early.
"I have the attention span of a strobe lit goldfish, please get to the point quickly!"
Yeah... "Leopard"... "Snow Leopard"... that's not gonna cause any confusion, right?
For the end user, it sounds like Snow Leopard is a minor upgrade. With bug fixes, performance enhancements, etc. It's a 10.5 -> 10.6 upgrade. Perhaps that's why they have a minor name change, from Leopard to Snow Leopard.
Or maybe they started following the Ubuntu naming Model. Let's see, is Hardy Hippo the same thing as Ubuntu 7.06 or what?
"Can of worms? The can is open... the worms are everywhere."
"That doesn't help with dual-boot PCs"
The GP was referring to a 'coporate' environment. It's pretty rare to have dual boot machines, it's either one or the other, with networked resources. If you want to dual boot, your data would still be stored on remote servers and accessed via CIFS/whatever in a corporate environment anyway.
I.O.U One Sig.
"The way the processor industry is going is to add more and more cores, but nobody knows how to program those things," he said. "I mean, two, yeah; four, not really; eight, forget it."
Hmm. Last I checked AIX and Solaris and HP-UX supported 128 CPUs or more. They all scale pretty well. So either he's talking about the desktop OS, where more than 2 CPUs is pretty new, or there's something new here.
I'd guess that it's not really something new; the basic problem of making programming for multiple CPUs "easy" has been around since the 1980s and it's still not "easy" -- oddly enough, you still have to think about concurrency, locks, multiple threads, etc.
Terrorist, bomb, al Qaeda, nuclear, yellowcake, kill, assassinate. Carnivore is dead... long live Echelon.
I'm sure many slashdotters have shared in the experience of a project rewrite that ended up bigger, buggier, and all around worse than the system or project it replaced...
Think about the cheaper and portable Macs though. The Mac mini and Macbooks could not have gone 64 bit immediately without increasing the size or heat output. And the tools in Xcode allow easy(ish) generation of Universal binaries which run on 32/64 bit Intel/PowerPC. I admit it's not as simple as pure 64 bit Intel, but it's not as bad as on Windows, where 64 bit adoption has been bad due to massive compatibility problems.
You see, this attitude of consumers is exactly why companies like Apple and Windows have so far focussed more on building OSes that look good, rather than work well. People want a shiny new thing, not a really efficient, rock solid operating system, because they have got used to crashes, useless error-messages, viruses and spam.
For me, this is the most enthralling idea in the End-User computer market in years. Finally, a company decides it's time to stop adding new eye-candy. Instead, Apple is taking a step back and taking their time to iron out the bugs and add actual innovation.
OpenCL sounds amazing. If it works as advertised, it will give developers who really care about performance the option to tap into the hugely parallel architecture available on the GPU that was inacessible to most of us so far (unless we wanted to learn the obscure proprietary semi-languages of ATI, IBM and nVidia).
Grand Central seems to be just the opposite of this: It will make sure those eight cores we'll soon all have in our machines will actually get used, even if the developers who wrote the programs we run didn't care to think about parallelization.
I'm bying Apple stocks. At a time when Microsoft's developers are once again falling victim to the marketing department (remember when Windows 7 was supposed to be a clean new start?), Apple is taking a bold step in what I think is the right direction.
...is solid as the Rock of Gibraltar on my MacBook. It's a stability improvement over 10.5.2 and a far cry from 10.5.0 and 10.5.1 which I avoided and stuck with 10.4.11. I'd put it right up there with Debian.
10.6 is something I'd be willing to pay for, though. Grand Central and true Intel 64 bitness would be awesome and make this MacBook rock. And as I mentioned earlier ZFS on a multi-disk future Time Capsule appliance would rock my world.
Knowledge is power. Knowledge shared is power multiplied.
I would mark you insightful, if I could. Moreover, if it really is a question of saving disk space by avoiding redundant different-architecture code, the installer should be able to do this just fine: Put code for both architectures on the install DVD and then let the installer select the right code for the machine.
There was no suitable 64-bit mobile platform when they made the switch (original Core Duo was 32-bit only).
When you sell as many portables as Apple does, that's an issue.
So the choice was either hobble along on the old-and-outclassed G4 for another year waiting for Intel (because there was just no way a G5 was ever going to shoehorn into a Powerbook), or endure a few years of mixed code.
Don't forget that 10.6 drops support for PowerPC CPUs!
The last Power Mac G5s were released in late 2005 and weren't replaced by the Mac Pro until late 2006.
The last revision to the PowerBook line was also released in late 2005. I'm still very happy running 10.5 on my 12" PowerBook G4/1.33Ghz from early 2004.
The last iBook came out in mid-2005, replaced in mid-2006. The last PowerPC iMac was released in late 2005. We have 10.5 happily running on my wife's 12" iBook G4/1GHz from 2003 as our kitchen TV.
It's pretty shitty that Apple is dropping support for machines less than 4 years old, and older machines that run 10.5 very well. It's especially galling that they are dropping support with a release that sounds like it should really be a free service pack or point release to 10.5 anyway.
obviously no deficiencies vs. no obvious deficiencies
If the things we value most in an OS are stability, performance and technical advancement, why are those the very things for which we are least willing to pay?
I've never understood why Mac nuts simultaneously claim that Macintosh is better because you don't need to replace your computer as often and do completely and utterly hate everything related to backwards-compatibility. It seems hypocritical.
Comment of the year
The success of the Wii owes much to Nintendo's brave (but wise) decision to persue a completely new customer base and leave the adolescent male (of all ages) market to MS and Sony.
The problem with the established PC/Mac market is that a big chunk of it have established skills and don't want (or don't think they want) a radical new GUI - they want a better way of running MS Office.
Its also worth wondering why the original Apple (after Xerox) GUI caught on. Now, I'm not going to dismiss all the psychology about desktop metaphors, but the big obvious factor that seems to get overlooked is simply this:
Before MacOS and Win3.1, if you wanted to (say) quit an application, it might be :q! or Ctrl-X-C or Ctrl-K-Q or Esc-X or /Q or /X or /E or QUIT or EXIT or BYE or. ESC and 9 from the menu or... Every fricking program was different. The IP wars of the time were not over software patents, they were over "look and feel" copyright of the basic menu structures.
After MacOS/Win3.1 it was File -> Exit. Ditto for Open, Save, Print... and the resulting dialogue boxes were all common, too. Instead of having to RTFM simply to find out how to open a file, everything worked the same way. It didn't matter if it was logically inconsistent to have "Exit" on the "File" menu you only had to find out once!
One problem now is we've drifted back to the application-specific GUI, as everybody invents their own system of dockable palettes, customizable tool bars, drawers, panes and other guff...
In a survey of 100 programmers, 111111 thought that duck-typing was a good idea.
I work in a multi-OS educational environment and see the weaknesses of all popular OS's in a short-exposure, high-contact learning context. The one area OS X really falls down is in the area of file-system and application navigation. I often see a student coming from Windows become comfortable managing both their files and applications with Linux (GNOME or KDE) far faster than they do with the Finder/OS X interface. While perhaps being a tired metaphor, the application tray, where any application minimised or otherwise can always be found (regardless of virtual desktop) works: they have per-application visual contact with what is active in their desktop session, uncomplicated by a dock doubling as a menu of popular applications.
After years of complaints from OS 9 and OS X users about the Finder Apple should confess to the difficult reality that - for many, not all - it is a major bottleneck to ease-of-use and therefore adoption. Students of mine - in general - spend far too much time second-guessing OS X where file and software management is concerned. Why are users' *losing* software and files so often that they need a *Finder*? Why are they so dependent on Spotlight that OS X might as well house all files in a flat-file-system? Why does the parent-window of an application still dominate the core navigation context even when minimised? This stuff confuses and frustrates people far too often I think.
It may not be the case for pro-users but I see students of mine spending far too much time clicking and dragging windows around in the course of trying to find and get stuff done on OS X.
My 2 clicks.
As the person responsible for a 4,400 desktop environment and as someone who deeply, deeply dislikes Microsoft, I can tell Apple in one sentence how to get Mac OS X into my environment....
Let me run it on non-Apple hardware.
I have a collection of Dell Optiplexes, HP dc7700 desktops, and a bunch of MPC 4x4 all-in-one systems. I would gladly, and with executive support I believe, pilot a Windows to OS X project on a few hundred systems within a quarter of that ability coming available.
Democrats and Republicans are like AIDS and Cancer, I want neither!
I'm an Apple admin (thank god) and talk like yours is getting really old.
Everyone has that one thing that keeps them from buying Apple products. ("real" video card in iMac, video camera on iPhone, etc.)
You already have an option. What's wrong with:
I have a collection of Dell Optiplexes, HP dc7700 desktops, and a bunch of MPC 4x4 all-in-one systems. I would gladly, and with executive support I believe, pilot a Windows to Linux project on a few hundred systems within a quarter of that ability coming available.
If you don't know what AltaVista is (was), get off my lawn.
I have a collection of Dell Optiplexes, HP dc7700 desktops, and a bunch of MPC 4x4 all-in-one systems. I would gladly, and with executive support I believe, pilot a Windows to OS X project on a few hundred systems within a quarter of that ability coming available.
Your problem is, like most people, you fail to realize that Apple is a hardware company.
Yes, they make OS X, and that's why people buy their hardware, but that's not where they make their profits. If they let people run OS X on non-Apple hardware, they'd go belly up very quickly.
> Translation: "We've put off any serious work on OS X for eleven months"
Pshaw. Means they're just done screwing with the interface for a while. They have a stable and useful user experience in 10.5.3. It'll get a few tweaks along the line, 10.5.3 changed Spaces considerably. They're also talking about major architectural changes to squeeze every last ounce of performance out of the hardware. You may not care about optimizing for multiple cores or offloading processing to the GPU, but the bioinformatics people who run racks full of Xservs in a compute farm were dancing in the aisles at WWDC.
>most likely will help get Macs into corporate environments
Licensing full Exchange support sure as hell will. The return of VB support in MS Office a year or so after 10.6 comes out will also help enormously. The Active Directory support keeps getting better and better every release too. With, again, more stuff licensed from Microsoft Apple will be able to play in the enterprise.
It's easier to be funny when you have a clue.
I'll give you the bug fixes though. Adding a new hardware platform did disrupt 10.5 and increased their bug rate as Apple tries to manage a common codebase for two very different platforms. Arguably, 10.5.3 represents where Leopard should have been at release, and could have been but for the iPhone. They're late, but catching up.
10.5.3 is full ready for use if you haven't switched yet, Check the remaining issues before committing though, there are (always) some bugs left.
Veteran, Bermuda Triangle Expeditionary Force, 1992-1951
I've got five Macs. My daily driver is an 8GB, 8-core Intel Mac Pro. My carry along a is loaded dual-core Macbook pro. Both are typically running linux, windows, and OSX all at once. I write graphics software for a living. Powerful graphics software, written at the metal level. I'm all for multicore/multiprocessor at the OS level; the easier, the better, and likewise, multi-machine for even bigger jobs. However, this does not change the fact that Apple is mostly doing iPhone work, and that not adding obvious consumer-level goodies to OS X will cost them dearly -- which they don't care about, because -- wait for it -- they're all about the iPhone now. I meant the post to be funny, all right, but only because it's true.
The very idea that low level improvements and bugfixes precludes feature addition at the GUI/high level is absurd, and if anyone at Apple had half a brain focused on the Mac, they'd never have said anything like that, or even implied it.
OS "features" can be as simple as adding a nice set of programs to the stable. Things like a decent personal finance manager. Wouldn't affect system stability one whit, but it'd increase the value of the Mac to the first time buyer by quite a bit. How about a nice, basic paint program? Or a set of kids coloring books / tools? A basic expert system? Lots of middle to high end users could use one, and heck, they're not that difficult to write. I wrote one in python that, minus the knowledge base, isn't even 10k and you'd be blinking amazed at how much it knows about rocks and minerals, and how well it can generalize and leap to conclusions. How about including a language teacher? How about a finder with a decent feature set? Something like... Pathfinder - buy it, maybe tweak it, and ship it. That would be @#$%^&*$ awesome. Heck, I'd probably pee right down my leg if they simply shipped a working, color version of midnight commander (a findery thing for shellfolk.)
See where I'm going here? Put an expert programmer in a corner, say "make a COOL one of these apps" and leave them be. In a year, if you don't have something really cool, the programmer should be shot. Total investment, one programmer's salary. Put ten programmers to ten tasks, watch em decently, and in a year, you'd have ten new selling points that had ZERO to do with OS stability, etc. Or just reach out the the Mac community and buy a few things, again, there are tons of them out there and I can assure you that many of them could be had for what amounts to peanuts. And also as we know, Apple's got more than peanuts in its pocket, and dropping a few million on programmers and/or acquisitions isn't a problem if they simply want to. So when they say "no features for you", what they're telling you is, "we're not going to exert ourselves on your behalf." They're not saying why... but just wake up and smell the iPhone marketing, man.
I've fallen off your lawn, and I can't get up.