Hidden Dimensions - The WWDC Keynote Technical Companion

by John Martellaro
August 4th, 2006

What the Agenda Tries to Achieve

Everyone likes to make predictions about the WWDC Keynote. That's the easy part. I'm going to do something a little different. My motivation is that while Steve covers a lot of ground, and it's a very entertaining show, very little of it is deeply technical. So I'm going to provide just a little more technical background that may help you put the Keynote into perspective and get more out of what Steve may say. And doesn't say.

The ultimate purpose of WWDC is to make Apple successful through the creative and dedicated efforts of Apple developers. As we saw with IBM's OS/2, an operating system can be technically elegant, but if no one writes software for the platform, it dies.

That idea alone generally drives the agenda of the WWDC Keynote. Even if not all the areas I discuss below are covered, what follows will hopefully provide some insight into what you'll ultimately see (or not see) on Monday.

1. Apple Success. All WWDC Keynotes have started out with a recap of Apple's recent financial successes. The goal is to get developers jazzed up about the business opportunity that's presented by Apple's hardware success. Also, there are analysts and reporters in the front row, Press section, and Steve wants to impress on them how smooth and successful the Intel transition has been. Look for plenty of comments about the wild sales success of the MacBook and Pro series with emphasis on the Switchers.

2. Partner Success. After establishing Apple's financial success, the next topic to cover is the intentions of the major partners and key applications that need to be converted to Universal Binaries. This is where Steve brings out the CEO of one (or more) of the major software companies to demonstrate their commitment to the Macintosh platform. The gist is that if the big guys are committed and making money, then get your own act together and make some dough as well! The number of Universal apps, in the thousands, will be touted.

3. Cross-platform Concerns. The next issue that has to be addressed is developer concern about customers running Windows on Intel Macs and abandoning Mac OS X. Charts and statistics will be presented to show that nothing of the kind is happening. This will be a good time to make some humorous comments about Vista so that the Apple developers continue their warm fuzzies that Mac OS X is the OS of the future while Vista is foundering. Boot Camp will be downplayed because Apple wants Mac OS X in your face all day, all the time. To emphasize this, Parallels may be demonstrated showing how, if you must, Windows can be dragged out for the occasional pie chart. This might be punctuated by a showing of the Apple Get A Mac "Work vs. Home" pie chart commercial to get the audience in a spirited mood.

4. Leopard Euphoria. Now that the audience has been assured that Apple is in phenomenal financial condition, that major software vendors are on track with Universal Binaries, and that writing software for Vista is for bozos, Steve will unveil a beta of Leopard and showcase some of the coolest features - features that Apple wants to show at this time. Other key features may remain cloaked until Vista shows more of its hand.

In order to show developers the huge payoff for implementing these new Leopard technologies, a demo by an senior engineer is typically called for in order to demonstrate how to use the latest version of X-code to light up the Leopard user's life with one of these dramatic technologies.

Some of the technologies that Apple started, like Spotlight and better file system metadata, will likely be in their second generation, and it'll be pointed out how several of these Tiger technologies, much improved in Leopard, compare to Vista. Again, the emphasis is always that writing software for Vista is not a worthy endeavor.

5. Professional Users. Apple's stated core strengths are at Mac At Work: creative, science, small business, and education. The first three of those, especially the creative markets, depend on fabulous desktop horsepower and graphics engines. But because professional users are very much also into MacBook Pros, and to create a sense of anticipation, the next generation MacBook Pro may be introduced first -- using the Core 2 Duo (Intel code name Merom). One of the keys to Core 2 Duo is the reduced power consumption and 64-bit hardware. Apple could choose to emphasize increased battery life because going from, say, 3 hours to 4+ hours is a lot more impressive than a 33% speed up in CPU power - especially since Apple's Intel positioning has been using numbers like 2x, 4x, 5x, etc. with the SPECfp_rate_base2000 and SPECint_rate_base2000 benchmarks.

Whether or not this Core 2 Duo-based MacBook Pro is announced, it's useful to remember that Mac OS X is a 32-bit OS. I haven't seen any indication that it's worthwhile to convert Mac OS X to 64-bits. Currently Apple suggests that if you need to run 64-bit applications that you factor your application so that the 64-bit computation part can present its results to Mac OS X's 32-bit GUI. This may change, but remember that Leopard still has to run on a lot of legacy 32-bit hardware (Power Mac G4s, PowerBooks, iBooks).

If the Core 2 Duo is introduced in the MacBook Pro, then the Mac Pro will probably be next on the agenda. The one thing you should keep in mind is that all recent Macs with more than one processor utilize what's called Symmetric Multi-Processing (SMP). This is a system whereby multiple, identical processors share the same memory. What's key here is that each processor has its own cache. If one processor's thread updates main memory, the other processor's cache is instantly outdated. Fixing that problem on the fly is called maintaining cache coherency. The hardware to do this for two processors, "2-way" SMP, and even 4-way SMP, in Apple's heretofore top end Power Mac G5, is intricate and expensive. I would be very surprised to see a desktop system with costly new Intel CPUs go higher than 4-way. It's a tradeoff.

6. Xserve. The Xserve is problematic. Apple has gone for a very long time without an Xserve refresh. The introduction of the new Intel Xeon CPUs begs for a new Xserve. However, there are some other considerations. First, the Xserve generally runs Mac OS X Server. There could be some timing issues with the Intel Xserve hardware and Leopard's features. Recall that Apple has never been reluctant to introduce new CPUs in the desktop system first where technical glitches aren't so mission critical. Also, there's no real practical or marketing need for the Xserve to be significantly faster than the desktop, at least in CPU horsepower. In fact, what drives their design is that the desktop systems can be liquid cooled (if necessary), while Apple keeps the Xserves air cooled. This is because, in a data center, if there's a fire, the last thing you want is exploding cooling systems spraying liquid. (Even if the liquid is electrically non-conductive like Fluorinert)

The Xserve/RAID is also due for a refresh. Everyone wants storage, and the Xserve/RAID is CPU independent -- it doesn't run a formal OS. Back when the Xserve/RAID was first introduced, Serial ATA (SATA) technology was still too primitive to put into a RAID storage unit whose hallmark is reliability. However, in 2006, SATA RAID is ready for prime time, so there might be a mention of a next generation Xserve/RAID moving from parallel ATA to SATA and even larger drives. A total of 10 terabytes would be a nice round number for advertising.

The Xserve and Xserve/RAID are exotic pieces of hardware used in business and science, and there isn't a lot of insanely great stuff to discuss. (Unless one is building a cluster.) Underneath the hood, there is sophisticated technology, but these systems are seldom fodder for a Stevenote. There is so much to rehearse for, these not-so-sexy products seldom make the cut because they make little contribution to the messages that must be conveyed in the Keynote.

7. iPhone. Everyone is guessing that the Apple iPhone will be introduced here. But I'll suggest that it will not because the Apple iPhone is more of an Apple opportunity than a developer opportunity. Unlike the iPod which produces music, needs speakers, earbuds, remotes, socks, FM transmitters, docks, and so on, the initial opportunity for Apple developers will be more limited with a phone.

The last thing Apple wants to do is erase all the good will they've painstakingly set up in the Keynote with developers and then finish by sending a negative message, "We've got our own cool hardware toy that competes with yours, will make Apple even richer, and by the way, there's not much opportunity for you to write software for it. Sorry."

8. iPod. No one knows if a video iPod with larger screen will be announced. We've heard only rumors about Apple selling, renting, or streaming theatrical movies. What I suspect is that politics with the movie studios is probably playing a larger and larger role. What the movie studios want to do is juggle the price of the movie product, its display mode, and delivery mechanism in a way that maximizes their revenue stream amidst massive global piracy. They want to protect all their traditional revenue streams such as cable, satellite, and DVDs. License concessions by the studios may dictate the design of Apple's next iPod.

There are oodles of issues to resolve. To name a few:

-- Steve was right when he said a few years ago that no one wants to watch a two hour movie on a five cm. screen. But what's changed is that people are catching up with missed episodes of TV shows, which are down to 42 minutes sans commercials, as they carpool to work. Or are watching music videos. Or snippets from YouTube.com. The design model for the iPod is shifting faster than Hollywood's decision cycle.

-- This unanticipated movement to portable video will certainly drive the market to bigger iPod displays. However, history has shown, with MP3s, that consumers are willing to sacrifice media quality for portability and quantity. If the iPod display gets just a little bigger, say 12 or 13 cm., it'll have the same apparent field-of-view, when hand held, as a typical sized TV. That presents a scary scenario of an affluent, mobile, reclusive, pre-occupied youth fixated on a single format, and that could threaten all the other legacy revenue streams.

-- Hollywood is clamping down on potential high definition piracy with HDCP protection and HDMI connectors on HDTVs -- which keeps the signal local. A new generation of HDTVs has arrived that accept and display 1080p. Blu-Ray outputs 1080p. Blu-Ray discs are expensive, and the technology is not yet assured success. Some have proposed that Blu-Ray and HD-DVD are already dead. No one yet understands how an expensive and restrictive HDTV environment's financial and market success will be influenced (or damaged) by a free-wheeling culture of low-def movies, TV, music videos and video blogs on a portable device like the iPod.

It's a lot of hard work to hash out the politics. That's probably why we haven't seen a more advanced video iPod and movie service from Apple sooner. If something is announced at WWDC, look beyond the initial hype. Check the technical specifications, limitations, caveats, license, connectors, interoperability, and so on to see how Apple has resolved these issues. If nothing is announced, you'll know why.

Hopefully, this technical background will give you some better insights as you enjoy watching or reading about the WWDC Keynote.