|HOME||PUBLIC LIBRARY||ANOTHER PERSPECTIVE||INFOPERSPECTIVES||CONTACT|
Lately, users of iSeries, System i, and Power Systems machines might wish their servers came from Apple, a company more famous than IBM for i stuff such as iPads, iPhones, iPods, iMacs. At least that's how investors see things. At the beginning of June, Apple's market capitalization was nearly $234 billion; IBM's market cap was slightly over $160 billion, about a third less. The companies may be quite different, but in at least one way their computers are similar: IBM's i and Apple's Mac are both firmware cuckoos, the former hatched in a Power nest, the latter raised in an X86 creche.
One might well be surprised by Apple's stratospheric market value. It mainly serves fickle consumers. Its rivals are the most vigorous competitors imaginable, such as Microsoft, Nokia, and Google. It is headed and very actively managed by Steve Jobs, who has suffered life-threatening medical conditions for the past few years.
Apple's boss fared far better than Randy Pausch, another creative force in computing who had at least two things in common with Jobs. One was a Whipple procedure. The other was admiration for work done in the computer science department at Carnegie Mellon University. Apple's operating system is a descendant of Mach, a CMU creation. Alice, a computer language descended from Smalltalk via Squeak, was shaped by Pausch at CMU (and funded by Sun Microsystems). Alice and related languages are an example of the kind of software IBM should be offering on i boxes to encourage a new generation of applications developers.
Apple has developed products that redefined computing, telephony, personal audio entertainment, and with its latest product, the iPad, it could well reshape digital communications. But Apple's success is due only in part to its hardware and software inventions. A lot of Apple's surge is the consequence of enlisting a brilliant army of application developers and, instead of paying them out of the Apple coffers, giving them development tools, providing them with support, and enabling them to be paid directly by application users. When the developers get paid, Apple gets a nice slice of the take.
This is more or less the opposite of how IBM behaves. Sure, IBM does offer some free software in the form of the Eclipse family development tools, but IBM charges a lot for its Power System machines and relatively more for the i 6.1 or i 7.1 systems software, forces developers to acquire (or rent access to) a server, and generally follows a business model that is a proven failure compared to Apple's. Notwithstanding the success of a number of OS/400 and i software developers a decade or two ago, the best efforts of the remaining software companies working the i field are simply not enough to spur growth.
If IBM needs some reasons to admire Apple, here are a few: While IBM, with annual revenue running at a pace of $97 billion, is nearly twice the size of Apple, with revenue running at $51 billion a year, Apple has under 35,000 employees compared to IBM's payroll of roughly 400,000. This translates into dramatically different figures for market capitalization per employee. Apple's share price makes the company worth $6.8 million per employee. IBM's share price pegs the corporation's value at $400,000 per employee.
Apple's annual revenue per employee is $1,490,000 and is vastly more than IBM's $243,000. No wonder IBM has so many staffers in Asia; that's the only place with wages low enough to fit the IBM model. Unless quite a bit changes inside Big Blue, as the economic tide rises across Asia, IBM will sink. And when it comes to corporate earnings per employee, the comparison is arguably even more stunning. Apple's shareholders, whose bidding has boosted (and skeptics might say inflated) its share price, still boast that their company brings to the bottom line more than $315,000 per employee. IBM reports annual after tax profit per employee that runs between $34,000 and $35,000.
While it's not too hard to explain or rationalize why Apple's numbers are better than IBM's right now, the size of the difference between investors' views of the two companies is still pretty hard to understand. But then so is the way IBM seems to have let the AS/400 lose its spark as well as its name.
The current Power Systems/i combo is heir to IBM's line of computers beginning with the System/38 that were based on the idea that a database management system and facilities to exploit it should be the heart of an information system. These days the database in an i box is called DB2 for i and it is truly a member of the database family that was born in the mainframe world and but which has subsequently been ported to every other environment IBM supports. Well, not quite. Application developers in some environments seem to have gotten more help from what passes for IBM's intelligentsia, the hotshots who contribute to IBM developerWorks. For years now IBM has given a free lightweight but nevertheless quite powerful version of DB2, Express-C, to anyone who wants to use it in Windows, Linux, or Mac OS. IBM has reached out beyond its own software in the Unix world, with a version for Solaris running on X64 platforms.
DB2 Express-C is not available for any of IBM's Power systems or its System z mainframes.
What are they thinking?
IBM is saying to the kinds of developers who have helped make the iPhone a stunning success and who promise to do the same for the iPad that the other guy's i, the IBM Power Systems-i, is not a place where they are welcome. Have in idea? Like DB2? Great. Sketch it in Windows, Linux, or Mac OS X. If it works you can distribute it free. If it needs a richer version of DB2 before it can emerge from its pupa, fine. Chances are the best price for DB2 will be available to somebody who buys an X64 box. And that server doesn't have to have an IBM label. It can be a Hewlett-Packard or Dell machine.
There was a time in the past when serious ISVs seemed to believe that IBM's policies helped protect them from raids by barbarian semi-pros. They have long since learned otherwise, although they might not dare talk about their observations to the IBM personnel they deal with. It is far easier for everyone outside Big Blue to just shine IBM on.
So as Steve Jobs committed years of his career to the kind of development process that began with larval Mach and ended up with the OS X big cat litter, IBM, like most of the computer industry, just watched and very possibly chuckled. Now that the computer companies Apple has passed by (in market cap terms) look chuckleheaded, Jobs has an option on the last laugh.
Meanwhile, development technology that could, if enthusiastically encouraged to grow, meet or beat what Apple offers has been kept alive, but barely so, at Carnegie Mellon. Alice in its current form is designed to help non-technical students build little object oriented applications in a three-dimensional space. It turns out that flow diagrams of business processes, like the crappy looking Alice bunny rabbits, may need three dimensions so that paths can cross. (If they don't, having an extra dimension or two isn't really so awful.) Could somebody (even an IBM software developer) put a friendly skin on IBM's DB-Freebie and find a way to bring it to Alice's world? I think so. If you think it's too hard for solo developers or small teams of intrepid hackers, it's time for you to visit an Apple store. Or download Alice and spend 15 minutes enjoying a starter tutorial.
And if IBM actually did find a way to mix common sense, open minds, and the practical demonstrations put on daily by Steve Jobs and the three-hundred-grand-profit-per-person creativity club called Apple, why should it do this with Power Systems running i?
IBM should get creative on its i for the same reasons Apple has chosen to keep its platforms (mainly software in the case of the Mac, hardware plus software in the case of the iPad, etc.) as Apple-ish as can be. The i platform is a truly IBM kind of platform. The platform's users are no longer tied to hardware, which gives IBM quite a bit of flexibility if it wanted to make some i platforms that are smaller and cheaper than what it presently offers.
Of course, Alice isn't the only possible starting point for a stimulating development environment, but it is a very good example because the kinds of languages IBM could use to get the attention of the developers it so desperately needs are the ones made for children and students. Anyone who has tried to learn a foreign language and discovered that the best reading materials for beginners are in the kids' sections of bookstores won't have to think too deeply to catch on. There are plenty of people who think Scratch is superior to Alice, some who still pray for Toontalk, and maybe even a few who think educators should have stuck with Logo.
What is clear is that making the i a great platform for wealthy team players who aced Course VI and wear propeller beanies is not going to get IBM the market cap its managers and shareholders want to wear.
— Hesh Wiener June 2010