As I flew into Boston, I looked over my schedule of meetings and events for MacWorld Expo. Near the top of the list, or actually down near the middle of the list but near the top of the nonparty items, were two meals that I was not looking forward to.
Wednesday breakfast: Listen to analysts discuss the true meaning of the Apple-IBM deal. Have to go because I promised Nancy.
Wednesday lunch: Listen to Roger Heinen, Apple vice-president and general manager of the software architectures group explain the true meaning of the Apple-IBM deal. Have to go on the off-chance that he might say something interesting.
Everybody wants to explain the true meaning of the Apple-IBM deal, just like everybody wants to invent the next Ben & Jerry's ice cream flavor. (Well, everybody I know. These are the people who invented the ice cream cure for the Grateful Dead concert munchies, Cherry Garcia. Ben & Jerry, that is, not my friends. My friends invented the snow cone drenched with Canadian Club whiskey called a Deadly Do-right.)
Although either meal would have been improved by a dish (or cone) of Ben & Jerry's, the food was OK. I did consider, when the moderator called for questions at the breakfast meeting, asking if there was any more coffee, but a waitperson arrived just then with a fresh pot, so my curiosity was satisfied.
The nonfood features of these meals were not that satisfying. I came away from each table hungry, neither meal having supplied the recommended daily allowance of true meaning.
Here, then, is the true meaning of the Apple-IBM deal. Are. Meanings. Plural.
But first, what it is.
IBM and Apple have agreed to agree to several things.
One of these is to work together to develop a multimedia platform and license it to others. This software platform is to include Apple's QuickTime, which defines a Movie data type and handles sound and video synchronization, and which is generally regarded as one of the hottest developments to date in the multimedia area.
This is significant, because the multimedia market looks like the computer and software industry's best opportunity to move beyond office equipment and into a real market: entertainment. Maybe "real" isn't the best word; "big" might be more apt. "Huge" would be apter. Replacing the typewriter is a fine thing, but replacing the TV and the VCR would be totally awesome.
Another of the hottest multimedia developments to date is the engagingly named video toaster, a nifty device based on a Commodore Amiga and allowing the savvy user to develop professional-looking videos with wild computer effects and titles and such for something closer to $10,000 than to $100,000. Todd Rundgren's video "Change Myself," which was available at the show, demonstrates dramatically what an artist can do with the product. The latest wrinkle is that the Toaster can now be controlled from a Macintosh computer. The toaster certainly has its limitations, but it is fair to say that this means one can now have a video production studio as a Macintosh peripheral for under $20,000.
What the toaster and QuickTime news means is that Apple has a large lead over Microsoft's Multimedia PC standard in this area, and that IBM is sharing in and further legitimizing the Mac approach to Multimedia. As analysts Tony Bove and Cheryl Rhodes have pointed out, the multimedia part of the Apple-IBM deal can pay off quickly, and can hurt Microsoft in this potentially huge market.
But it is the other things that Apple and IBM agreed to agree to that are getting most of the attention, because they threaten to redefine the computer industry as we know it, rather than just to create a new, larger industry. These things have to do with architectures and operating systems.
Apple will build future machines using IBM's RISC architecture. Apparently Apple's own RISC development will be dumped or backburnered once Apple fully commits to the IBM chip. Given Apple's history of keeping its options open, it probably won't get backburnered. Through his brother Peter, who was at the show, I heard that Apple veteran Allen Baum, who has invested some time in Apple's RISC technology, finds the IBM approach technologically unimpressive.
IBM will incorporate the Mac Toolbox into its Unix version, AIX, so that it can run Mac applications. That doesn't make a lot of sense, I know, and there seems to be some disagreement between Apple and IBM about just what AIX will get. IBM has said that AIX will have the Mac interface, and Apple has said no. I think there may be some details to be ironed out here.
IBM will integrate Macintosh into the IBM client/server architecture. Apple will get legitimized in corporate America, is the general view of the significance of this.
And finally, the two companies will get together to develop a new operating system. This is specifically what everyone wants to talk about, and what my Macworld Expo meals featured as the main course.
The new operating system is to be based on Apple's Pink project, is to be object-based rather than file-based, is to have a graphical user interface, is to be portable, and is to run across networks. There are a couple of levels here: Underlying everything is to be an object layer, which supports the Pink operating system, API, and GUI, as well as supporting existing operating systems, meaning some kind of Unix, OS/2, and the Mac OS. Everyone seems to want to call it all Pink, although Bove and Rhodes sort of suggest that when IBM has finished making its contribution the result will be Purple. At the breakfast meeting, Macromind founder Marc Canter expressed his view of what the software will look like when IBM gets done in language of the sort that any day now the Supreme Court will probably decide is not Constitutionally protected. Marc Canter has a lot of opinions. He would probably like to rename Ben & Jerry.
The two companies will form a jointly owned company to produce this new operating system, transferring their respective object-based operating system groups (about 100 employees each) to the company. Basically, Apple's contribution is Pink, and IBM's is Metaphor.
Metaphor is David Liddle's company, which last year formed Patriot Partners with IBM to produce platform portable object-based system software. Liddle's vision was something called Constellation, which was to be a tool-rich environment oriented toward user tasks rather than being defined by discrete application programs. IBM has since bought Metaphor and is delivering it and Liddle to the new venture; Patriot Partners and Constellation have ceased to exist as separate entities, and Constellation is supposed to become part of the new thing.
Just how Pink and Constellation get integrated is unclear, but IBM's idea was that Liddle would be in charge. It looks as of this writing like Liddle will be the president of the new company, but Apple is wrangling to maintain some control over the venture, pushing for Ed Berst to be the CEO. Apple and IBM are to own 50 percent each of the company.
That's the deal. Not surprisingly, it is unpopular among two groups of computer users: those who have long admired Apple and those who have long distrusted IBM. During the show, you could buy a t-shirt on the street in Boston that expressed their common view: On the front, the Apple logo and the caption: This is your brain.
On the back, the Apple logo done in IBM-blue stripes and the caption: This is your brain on drugs.
Opinions vary on when we will see the new system software.
At the breakfast meeting, pundit Jeffrey Tarter gave his prediction: five years from now at the earliest, if at all. This sounds right, but one hears claims that something could appear in the next year. T/Maker founder Heidi Roizen cleared that up by pointing out that there was confusion in the press about when a product might ship and when developers might get a first look. Roger Heinen predicted that the Mac operating system would be running on RS/6000 RISC machines in 18 to 24 months, presumably for developers to see. There are predictions that AIX will be running Mac software by 1993 and that new RISC-based Macs will be on sale in 1993. And Heinen has predicted that the Pink market could be where the Mac market is today by 1995. This is surely hyperbole unless Pink is almost ready to go out the door now, and in any case would seem to be inconsistent with his prediction that the Mac market has another good decade.
How much pain will be involved? Well, Marc Canter says that Pink FUD (Fear, Uncertainty, and Doubt) is holding up software development, but Heinen insisted over lunch that Apple was committed to binary compatibility for existing Mac software on new machines and under Pink. Unix support is more confusing. Apparently Apple will continue to support its Unix, A/UX, on its hardware, provide Mac Toolbox support for IBM's Unix, AIX, and evolve A/UX toward AIX. In the long term, though, AIX is likely to be supplanted by this thing called Pink.
Who wins and who loses? This is one of those questions that journalists like to pose, to the frustration of those unimaginative enough to be looking for answers. (The reason for this need not concern us here. Oh, all right; it's because most journalists are philosophy majors, because journalism school graduates can't spell.) I certainly don't know the answer, but here are some cautions regarding some speculative answers being circulated.
Over seafood out by the bay the last night of the show, two Fortran developers confessed that they were asked not long ago to do a Cobol implementation. One of their reasons for refusing, they explained, was that "nobody programs in Cobol." It turns out that they weren't just commenting on the obvious decline in popularity and fashionability of Cobol; they meant that even those who use Cobol do their development using other tools.
Cobol defenders often cite the 80 Godzillion lines of Cobol code now in use, and while it is easy to take potshots at these references--like saying that 80 Godzillion lines of Cobol translates into 3,000 lines of C--there is a huge body of Cobol code in use, and this argues for Cobol's being around for some time to come.
But if nobody actually programs in a language, doesn't that make it officially dead as a programming language? Even if we can't pull the plug on Cobol, can't we agree on the diagnosis of brain death? And in deference to the 80 Godzillion lines of code, start referring to Cobol as a code repository? It's just a thought.
Dave Winer was at the show demonstrating the latest beta of Userland Frontier, his system-level scripting tool for the Mac. To make the product less intimidating to users who might be intimidated by the idea of programming, Winer's people have given Frontier the ability to spit out scripts as small double-clickable files. Formerly, the scripts lived in Frontier's object database. Double-clickable Frontier scripts are not stand-alone applications, but Frontier documents, requiring Frontier to run. Their advantage lies in their reassuring concreteness, and in the possibility of sharing, trading, and downloading them.
The downloading angle could have something to say about the success or failure of Frontier. If lots of Frontier scripts start appearing on electronic services free for the downloading, it could trigger a lot of purchases of Frontier. On the other hand, if those scripts are bad, they will reflect badly on Frontier itself, as happened with the glut of bad stacks when HyperCard first came out.
I've heard this analogy to HyperCard stacks, but it seems to me that it isn't very good. HyperCard stacks are usually judged to be bad because of design issues: They are visually awful, they are poorly organized, it's hard to tell what they do or where you are in them. Frontier scripts will have none of these problems. Not being visual and not having a structure saves them from problems of visual design and structure.
Of course, they may have other problems, like crashing your system or trashing your files. Frontier's language is powerful and reaches deeply into the system or an application. Frontier scripts can do damage. Scripts from amateur programmers should be examined or tested before being accepted by sysops and before being used by downloaders.
And the door is now open for amateurs to do the kind of deliberate damage that virus creators do. That's the price of power.
As I look over the press coverage after the show, it appears that the favorite flavor of speculation is that IBM will eventually buy Apple. I'll make a specific prediction about that: It just won't happen. Feel free to ridicule me when it does.
Copyright © 1991, Dr. Dobb's JournalThe Toaster is Hot
Purple Prose
Your Brain on Drugs
Meanwhile, Back at the Show