Slashdot Mirror
Apple OS X, BSD and Jordan Hubbard
We've had a number of posts noting that Boston.com's digitalMASS has a very decent article on Apple's OS X, BSD and Jordan Hubbard.
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
seems hard to believe that he had to struggle to land the job at Apple, as such a prominent OS developer. I would have thought that the more successful/visible open source developers would have their pick of jobs at any firm... and Hubbard would be especially well suited to work on OSX, since it's based on freebsd. i bet he's just being modest...
I've been a linux user for almost 6 years, and generally laughed at Macs (the whole one-button thing, etc.)
I wanted to find a nice laptop that would run linux, and even had a dell for a few days. Then, a friend of mine introduced me to Macs, and, in particular, the Tibook. You can see it yourself - overall, its probably the single best piece of hardware engineering imaginable.
And OS X really is awesome. I'm not into having the point-and-click interface myself, and love the console. But OS X really is nice to use. Its networking support is amazing, and works right out of the box. Support for sleep is great too.
Right now, from what I can see, the biggest problem with OS X is the lack of a decent DivX player. (4.11 tends to desync in about a second). Otherwise, it's awesome. And, if you really can't let go of blackbox or whatever (like me), there's the XDarwin project that lets you run X on top of OS X. So far, I've only tested the default twm, which runs fine. But using the apple developer tools you can compile any window that's been ported (I believe at least gnome and afterstep have been), and run it there.
Certain products are still not quite ready for OS X, but the situation is improving rapidly. I have to disagree with one of the posts below - its not about being "productive"; one could easily do that in Linux. (I refused to run IE, and will
NOT be getting Office). But it is a sincerely nice operating system to use, and the hardware is definitely going to be a computer legend.
Regards,
trurl
Essentially, I spent the last ten years of my life shackled to Microsoft products with the all-too-infrequent practical use of Linux. As Microsoft's business practices continue to get ever more predatory, and the Microsoft operating systems become increasingly marketing tools rather than productivity tools, I decided that it was about time to try something new.
I found an inexpensive, new iBook, and bought it. An "icebook" with a 500Mhz G3 processor, I've been quite happy with it so far. The construction of the iBook is quite decent, with a few common blemishes in the casing and a few mechanical defects reported. However, the real shining star of Apple's lineup has got to be OS X. This BSD alteration (Or enhancement, or bastardization, or annexation, call it what you will.) is positioned in the perfect place to bring intelligence back into the use of personal computers. Functionally, OS X is a wonderfully complex yet artistically presented program interface which does an admirable job of concealing the true nature of things from the average Macintosh transitional user, while providing an extremely high amount of flexibility for the more technically oriented. With the Macintosh userbase, there's actually a very devoted core that could use the help and assitance of open source efforts despite the problems with Apple in regards to certain areas of the system. (The interface, primarily)
Projects suck as Fink, an excellent tool for porting unix applications to the OS X environment are a great start, but what will really help Apple prove a real challenge to Microsoft is the conscious effort by Open Source developers to port applications to Apple hardware so seamlessly, that the average user won't even have to know that The Gimp was actually a unix application.
This is where Apple has succeeded as a core business, making computing simpler for the artistically, rather than the technically minded. The best thing Open Source can do is aid the Apple userbase in proving that the Mac is a viable alternative. Yes, Linux and BSD themselves as well as all the other systems out there, deserve to continue to be the primary focus of most efforts. But it just may be that the most effective way to open up the operating systems market will be to back the entrenched underdog.
My own pointless vanity vintage computing page
Well actually, in Australia at least (and presumably in the US) we have this little law against selective punishment which states that if you are to take legal action against someone you must also take legal action against any other infrignement of the same magnitude against the right you are protecting. In other words, if you regularly let people you don't know walk through your property as a short cut then you cannot sue a specific person for taking the same short cut without showing that you are also going to persecute all future infringements. Otherwise the legal action is discriminatory and will be thrown out of court (see the case of Williams vs Thorsbourne in relation to the Port Hinchinbrook Project). Apply this to copyrights and you will find that they are only viable to the extent that they are protected.
Now, even if this law does not exist in America, the fact that it does in Australia would be enough to convince most international companies to be strict about enforcing their trademarks to save trouble in the long run. Of course, ianal...
Regardless of all this, Apple is certainly not a highly proprietary company any more. Lets take a look at some of the things they work with and sell these days and how open they are:
In fact there is very little that Apple do that is proprietary anymore. So they defend their look and feel vigourously because that's about the biggest thing that sets them apart from what others could provide. Almost everything else they do is opensource or follows widely published standards.
All in all, Apple looks like a pretty good place for an opensource advocate to find a home.
I use FreeBSD at work. It's awesome. I also use OS X daily at home and at work. And really, I can't see much difference between the Darwin core (pure opensource) and FreeBSD in terms of out-of-the-box functionality, except that several of the utilities and bsd apis are dated (no localtime_r, for example). That's not a big deal really, it's Apple's goal to keep Darwin in sync with FreeBSD, and anything that you are missing (as an application developer), there's not too much stopping you from getting sources and compiling it yourself.
My point is, Apple is shipping a fully capable open source OS. I'm talking about Darwin, not OS X. This is what Jordan is working on. If anything, it aims to succeed FreeBSD. FreeBSD is nice and all, but it's got everything that makes a modern Unix, which is exactly what keeps its market penetration small. It doesn't have the attention that Linux has to keep it growing.
The difference that Darwin has is the incredible number of NEW ideas that make a NEW OS. Darwin isn't just another BSD. Apple has learned a lot having several OS projects fail (you have to admit, you do learn a lot that way), and the good stuff has gotten into the Darwin core.
There are things like XML and Unicode support at the base level of the OS. XML is used to describe everything in the sytem, it's basically the conf file format of OS X for everything that doesn't have too many legacy dependancies (for example, fstab is still there, but from what I can tell, it isn't the source of the information at boot time- NetInfo generates it. mount et. al will use fstab still). High-level, object-oriented frameworks are available with at the base level of the OS through CoreFoundation. CoreFoundation is basically the core of the OpenStep Foundation framework without Objective-C. It's pure C. Everything in the core of OS X is written with CoreFoundation- it provides the XML and Unicode support (this means kernel modules, etc). Every string you use is a Unicode-capable string. The powerful array and dictionary 'classes' you use can be converted to and from XML (simple data archiving and unarchving, and human readable too). And you can download and install CoreFoundation onto Linux/FreeBSD if you want to.
I can hear people in the background yelling 'OMG! They're butchering BSD!' No, their upgrading and replacing some very old systems with new ones. They're replacing them with things that fit the needs of modern workstations, not just servers. Things like 'ls' and 'more' aren't touched. They're user utilities, and pretty irrelevant at the OS level (and fairly irrelevant for Apple anyway, because a cool ls isn't too useful in OS X).
You can check out Darwin yourselves from www.opensource.apple.com. You'd probably have more fun with Linux/FreeBSD for a few years, though. While Darwin doesn't have everything it needs to make a long-time *nix user happy, it does have everything it needs to bring all the power of Unix to the masses.
Even before Rhapsody was remade into Mac OS X, I remember Mac OS Rumors article about a rumor (surprise surprise) that what is now Darwin would be available free to universities, simply because Avie Tevanian wanted to give back to the community (as the BSD license doesn't force you). Well, they gave everything back to everyone, and then some.
I'm not saying that Apple is all good, but so many people really underestimate what Darwin is and what it stands for.
Moderators should have to take a reading comprehension test.
I do not mean to start a flame-war, and hope this post will come up merely as an opposing viewpoint, but:
GNOME and KDE do not provide an end-user experience until the end-user has already gotten past the usual linux hurdles -- getting X to work with their graphics card, configuring their network and so on. This is trivial for a lot of us, but it is not for my standard of maturity for a user-interface ready for general deployment: "Can your mom set it up without your help?"
The answer, generally speaking, is no.
MacOS X is the first counter-argument to "Unix is not for peasants" that is pretty much true across the board. Your mom can install it and use it. It's really friendly, and it's really effective.
It's redundant to point out that the Aqua interface and the MacOS Finder are simply a candy-coated veil covering up a very mature and stable bsd/unix environment appropriate for the same range of tasks to which desktop linux distributions are currently applied.
Whether or not KDE and Gnome, or the open source movement as a whole is "thinking about the end user" is a moot point, but that these things are not ready for general distribution to Your Mother is pretty much inarguable.
As a user of both types of systems, I can say that OS X has provided me with the best user experience since the first time I sat down in front of a NeXTstep system.
Something I've barely seen mentioned heretofore is that MacOS is not really new, so much as it is a complete overhaul and Apple-ification fo NeXTstep. The NeXTstep user experience was unparalelled at the time, and I'm glad to have it back in a thoroughly modern form with such a magnificent GUI.
As I see it, an open-source base OS is apple offering a laurel wreath to the open source community and extending a modern standard to everyone (why else would they release an intel version themselves?) that can be freely used without the commercial side of the product. However, if you're willing to fork over the money / buy a macintosh, you're in for one serious treat -- probably the best user experience you'll ever have.
I would like to see someone light a fire under apple's butt to get a few details straightened out like a better software sound subsystem and support for the peripherals traditionally associated with the apple market -- like scanners (AHEM!) and the Soundblaster, and I'd like them to return to providing onboard audio input so i didn't have to talk into what approximates digital soap-on-a-rope, but I wouldn't switch to anything else despite these issues (And the basic support i need is still available in OS X through the classic environment, so a lot of these concerns are taken care of in a temporary way a the time being) and I would never be able to sit down in front of GNOME/KDE with a straight face and say, "this is ready for the market! woopee!"
-JSJ
In effect, all Linux proper is is an OS kernel. Everything on top of the kernel is something that is bolted on independently of any kernel development. Thus Slackware is the Linux kernel plus "all sorts of stuff Patrick Volkerding added;" Red Hat Linux is the kernel plus "all sorts of stuff they added;" ditto for SuSE, Debian, Mandrake, ad infinitum.
With the BSDs, there's quite a lot of additional "environment" that is tightly tied to kernel development so that you've got a "base system" that is defacto-standardized that is capable of, for instance, recompiling itself.
With Linux, you've got to add in whatever that is needed that isn't in the kernel in order to do that yourself.
With that larger basis of "stuff" surrounding the kernel, a whole lot of the arguments "Red Hat puts the files here; Debian puts them there" just plain go away. The "Linux Standard Base" effort where they're trying to standardize where a bunch of the basic stuff goes and what it does is an effort that would be ludicrously irrelevant amongst the BSD folk; they started off by standardizing the user space stuff that LSB is fighting over.
Then there's Ports. Ports is sort of the BSD equivalent to Debian's apt-get or perhaps the Red Hat-oriented autoRPM . Except with a difference: With Ports, the approach is not to download binary packages, it is rather to download the sources, pull in any patches needed for Ports integration, and then compile it all.
That's got the demerit that it's a lot more work for your poor, overworked CPU.
However, it has the merit that if you compile libraries and packages, together, on your system, with the same compiler, the sorts of "DLL Hell" that people suffer from when they grab RPM files from here and there just can't happen. The libraries will necessarily be compatible with the applications because the applications were compiled with and for the precise set of libraries you have on your system.
This means that if there are any challenges in getting programs to compile, you'll hit them. That being said, since the folks collecting and maintaining the Ports will indeed hit those issues, they're likely to have patches in place so that by the time you see the code, it should compile cleanly.
In effect, the crucial concept involved in all of this is that the BSD packaging paradigm is that everything should be readily compilable and recompilable, from the ground up. The classic "make world" will compile all the base tools, the kernel, all the kings horses, and all the kings men, and what you get at the end is that every single component in the "world" (which is the base system; the stuff below Ports ) has been rebuilt locally.
It's all using Makefiles, and can be downloaded using CVS, so the details are all visible. None of the controversies of "well, the Red Hat kernel compiles include some special patches, and getting at them is a bit challenging...."
Big-time learning opportunity.
If you're not part of the solution, you're part of the precipitate.
First and foremost, we must consider the interface. Here we are talking about OS X/XP vs KDE/GNOME. If you have used all four you can attest to the fact that KDE/GNOME have come a very long way, but are still very far behind, and if we strictly talk about KDE/GNOME vs OS X (since the play-doh theme in XP has shaken what faith I had in MS' interfaces), you must admit that the open source desktop environments are 2-3 years behind. Now what troubles me more is that readers are in denial about this, and this lack of understanding about what the experience needs to be stemming from the fact that OSS OSes are used primarily by programmers/admins/etc. prevents open source desktop environments from competing. Even you say,
KDE already deliver all the stuff Jordan Hubbard was talking about, even before OSX was on the shelves.
I hope this is not meant to insinuate that the KDE experience is comparable to the OS X experience. I actually read a post that said that Apple should, "port Aqua to X windows". If you think that you can run the Aqua interface on top of XFree86 (or one with comparable features, not just a bunch of pretty pixmaps, which is what the Mozilla organization seems to thing Aqua is, most unfortunately for those of us who want to use Mozilla for OS X), the future of OSS desktops is doomed.
Now, while I find the "open source is doomed forever" attitude unfair, if we take a look at where desktop functinality is right now, open source has lost. As I said, KDE and GNOME are not even competitive with OS X, the GIMP is nowhere near being competitive with Photoshop, nothing is competitive with Final Cut Pro or Premier/After Effects, nor are there substitues for the iApps (simple, but still extremely funcitonal consumer-level apps), there are very few games brought to open source operating systems, although Apple has a problem with this too, they manage to get a port of virtually all the top-shelf games, apps like Maya that used to be the domain of UNIX-like OSes are now on OS X, eliminating the need for a Photoshop Mac and a Maya SGI on your desk, and finally I must say that open source office products are competitive with MS Office, but must also admit that Office v.X is truly a very powerful suite and the best availible tool, although still only worth a fraction of its $500 price tag.
So to summarize my points, open source software for the desktop is currently not in the same league with the commercial software, but it could get there if more effort was focused on it, and it is completely reasonable for Hubbard to go with Apple and focus solely on making the best possible software, as open source solutions, even though they may become an extremely viable 3rd desktop platform one day, probably will never reach an elegance of interface of Apple products.
"Reality is just a convenient measure of complexity" -Alvy Ray Smith