And same goes for this Panther release. This release of the OS -- at least the features presented to us -- did not reveal any major innovations. No re-definitions of how we do things. Please, don't think that I am negative on Apple, because it isn't true. But I do see things the way they are and not as Steve Jobs would like me to. Last year's Jaguar WAS innovative, WAS revolutionary: Quartz Extreme, Renderzvous, Inkwell, Sherlock 3. These were brand new features not found on other systems of the time. With Panther, we actually see Apple try to USE these technologies into their system instead of just having them lurking in the background having us pray that new apps might come out that actually use them. Yes people, Panther is the continuation of Jaguar; it is the version of the OS that uses these last year's innovative technologies, but it doesn't really include any new innovations of its own (except being 64-bit that is, which is not really a user feature, but a system feature). Where is full MIME support for example? But hey, you can't (truly) innovate every year, I'll give you that. ;-)
Then, it was the time for the G5. Yes, this is an exciting hardware release. The G5 is a solid and fast machine. Some people said that it doesn't look sexy, but I think that we see a turn of Apple into more workstation markets, so this "professional and serious-looking" case is actually appropriate and to be expected.
So, Apple now has a brand new 64-bit processor, the PPC970, with up to 2 GHz speeds and promises for 3 GHz in 12 months. Support for FW800, USB 2, AGP Pro 8x, PCI and PCI-X (depends which model you buy), 9-fan but quiet case, up to 8 GB RAM ("broke the 4 GB barrier" as Steve Jobs puts it), 1 GHz bus for the high-end model.
Let me jump the gun and say that if you plan to buy a G5, get the dual 2 GHz one, yes, the $3000 high-end one. The lower end ones are bad buys (especially the 1.6 GHz model). I don't know what Apple was thinking when creating the 1.6 GHz (money probably) but they use DDR333 on a 800 Mhz bus. That means that the CPU has to wait for the RAM to finish its cycle before it receives the new data. They should have included DDR 400 on that model too in my opinion.
Apple showed us demos of Logic, Photoshop, Mathematica and a 3D rendering app going against a Dell dual Xeon machine. We were not told of the specifics of the Dell Xeon machine, nor if HT was properly turned on in the BIOS. We were also not told if these Mac apps were specifically optimized for the G5, e.g. if they were versions that will never see the light of day on a retail box, but built specifically for the demo. Anyhow, on all instances, the dual 2 GHz G5 had much-much better performance than the Dell machine. However, I would advise to not jump the gun so fast on this. I mean, come on, iTunes and Safari are still not baby-smooth... when resizing their windows on the fastest G5! I played with the machine for a few minutes and was [again this year] negatively surprised by this simple thing that Apple still haven't being able to master with both Panther and the new CPU: scrolling and resizing. I think I will send a copy of BeOS or Windows XP to Steve Jobs for Christmas, just so he can compare.
Additionally, Apple did the SPEC benchmarks using GCC 3.3 on both x86 and PPC, while the vast majority of the C/C++ developers in the x86 Windows land actually use the much faster and much more optimized for P4/Xeons/HT Intel ICC compiler. Then, you will probably find out that Apple's numbers are not really that fair. They keep calling their x86 benchmark results to have come out of the "best PC money can buy," but they don't explain to us how the Pentium 4 scores more than ~1200 at SPEC.org, while Apple gives it a measly ~800 number on its own results. Apparently, the PC they tested with, it's not the fastest P4 money can buy. Most user-oriented applications are using integer and not floating point anyway (PPC970's main strength). But workstation-class applications might need fp. [Update: Others had similar thoughts as well regarding the real speed of P4/Xeons. Here is another article as well at hand. And another one.]
I am not saying that G5 is slow. It is not! It is a fast machine. It is the product that has made Apple really caught up with the competition. But I don't see the dual G5 at 2 GHz overcoming the x86 today. Intel released today a 3.2 GHz P4 and they expect a new P4 version (faster core per same speed), to go all the way up to 4 GHz before the end of this year. Apple's roadmap is to reach 3 GHz in a year from now. So, has Apple caught up now for real, or Intel will speed through again and leave Apple in the same condition as it was until last Sunday night? We will know in a few months.
My other problem with the G5 is its entry-point price. The lowest-end model (the one with the slaughtered memory version) starts at a minimum of $2000. Seeing how close the prices of the 1.6 and 1.8 are, I would argue that Apple could go lower and introduce the low-end model for $1700 or $1800. This new G5 business, are really in need of a cheap offering. Oh, and while you are at it, move down the price of the eMac G4 to $499 to better compete with the cheap PCs. Apple really needs to offer cheaper computers. In my opinion, the important thing in this point in time in Apple's history is not the profit margin. It is the market share percentage. This is where Apple loses a small sum everyday these days and it could prove fatal if developers leave the platform because of the declined user base. So Apple, show us that $499 eMac of yours! We know you got margins!
Anyway, enough rambling. After the keynote was over, we got downstairs to the actual showfloor, and it was nice to see companies like Perforce, Oracle, OpenOffice.org, Metrowerks, Trolltech, Big Nerd Ranch, LaCie, O'Reilly, 4D, Frontbase, REALBasic and more. The participation of the Mac developers was quite high. There were many people at the conference, even with this hefty registration fee. I suggest you go over to the conference sometime this week if you got the needed money to register. It is a nice, interesting, well-organized event overall with some interesting development sessions throughout the week. Get yourself up to speed with the Apple world. It's worth it. Just don't take into account everything as served to you by marketing, but use your head.
- "Innovation or Catch Up, Page 1"
- "Innovation or Catch Up, Page 2"
