That pretty much describes the MacHack Technical Conference, which just had its ninth annual conclave in Ann Arbor, Michigan on Wednesday June 22nd through Saturday June 25th. This was my fourth time in attendance, and I can safely say that I have never attended any other business meeting that even came close to MacHack for the sheer volume of information discussed, or for sheer fun.
MacHack '94 started with a keynote address on Wednesday night at midnight. This timing gives an important first clue that MacHack is not your "normal" business conference. Another clue was the introductory notes from the conference organizers including such advice as, "The hotel says that if you want your room cleaned, the `Do Not Disturb' sign must be off your doorknob by 2:30 pm. Also, you'll be glad to know that our contract with the hotel says that they will not attempt to make up your room before noon."
Although the focus is Macintosh, MacHack is in no way affiliated with Apple. That makes for a much more honest and open conference than Apple's own annual Worldwide Developers Conference, which tends to be tightly controlled.
That concentration of talent is one of the great joys about MacHack. Every year, I save up some of my toughest questions for Ann Arbor. For every conceivable Mac question, I know there will be a dozen people who can provide answers. Better than that, I know I'll have a good chance of hanging around with someone who actually was involved in creating the system or program involved.
This is a gender-lopsided group: I counted no more than a dozen women among the 300 attendees. Interestingly, the first MacHack I attended had a similar ratio, but the conference organizer, the conference chair and the project leader of System 7 (the major topic at that year's event) were all women.
Social or fashion skills are notably not major criteria to attend MacHack: standard dress is a T-shirt and jeans, unkempt hair and sneakers. There were guys with no shoes, guys with long fingernails, and lots of pizza-bellies (beer is not a passion with these guys; Mountain Dew is dominantly the beverage of choice).
For these three days, the Machine Room becomes the world center for fever-pitch Macintosh tinkering. Questions are shouted out loud (and answered in the same way from several places around the room), tables are covered with circuitry and tools, there is frantic snooping at each other's projects. Screens flash in odd ways, strange sounds emerge from unlikely devices, all while electrons are coerced into doing the darnedest things. It's nerd nirvana.
A major issue in this session (and in several other sessions) was "Chicago", the code name for version 4 of Windows, which will be released early in 1995. Important for Mac fans, Win 4 will include two elements that are key Macintosh advantages right now: a much simpler interface (deep-sixing the goofy Program Manager/File Manager dichotomy of the current Windows), and the potential for simple attachment of peripherals ("Plug and Play"). Plamandon said that the Win 4 interface borrows some of the best ideas from a variety of sources, saying, "If you use one source, it's plagiarism. If you use many sources, it's research." He also reported that Windows has sold 50 million copies, and currently adds a million a month. He added the scary thought, "If Apple couldn't successfully market Mac against DOS [said with a guttural sound of disgust], then how do they expect to compete against Chicago?"
That's not quite a fair comparison, of course, since Apple never tried to market the Mac operating system on its own merit - a potential buyer always had to buy a whole computer from Apple just to get the wonderful OS. Finally, last fall, Apple announced that it would now offer the Mac OS for licensing. But, as we've all noticed, nothing has happened since then. Plamandon suspects that Apple management is in a state of shock over lack of manufacturer interest about its belated offer, after many years of being begged to license the OS. Indeed, Compaq has not only turned Apple down, but has ridiculed Apple in the press.
Plamandon described the Intel chip architecture (the heart of DOS PC's) as "grotesque" and "bizarre", a surprising choice of words for a Microsoft guy. But he said that was the price for Microsoft's paramount design principle, step-by-step migration of its customers to improving technologies.
Plamandon reported that Mac programmers almost universally hate the Windows API (method for programming) because of its byzantine architecture, but suggested that Mac'ers would feel more at home with Win32s (if you're interested in the wonky details, call me).
Finally, Plamandon decried the "ugly" style of many existing Windows products, and said that (I'm quoting) "Mac programmers have a much higher standard of code quality than Windows programmers", and that he would welcome the application of those skills to Windows.
By the way, I glommed onto a stack of ever-so-fashionable System 7.5 temporary tattoos. So please call me if you'd like to emblazon your chest with a picture of Mozart (the development name for the system) and the legend, "Born to Code System 7.5."
All of which made the sessions on Apple Guide and System 7.5 quite interesting. We learned quite a bit about how to employ these exciting new technologies. I also had the opportunity to spend time one-on-one with the project manager for Apple Guide for some terrific conversation.
A favorite anecdote about naive users in need of help: One neophyte was learning the Finder and was told to use the "Special" menu, and said, "Which menu is `Special'? - they all look the same to me."
Well, it turned out quite nicely. I got a decent-sized audience, with great questions and comments, and the right amount of applause upon conclusion. And they laughed at my jokes at the right times.
Best of all, two heavy-duty Mac stars who I admire were in the audience at my dog-and-pony, and they both said nice things about what I had done.
A note to Dale Coons of Lintas: one speaker mentioned that Word Perfect's own internal crossplatform tools (which Dale's brother helped to design) were done "exquisitely well".
One of the NDA sessions laid out Apple's plans for future development of the Mac OS. I think most of the attendees were pleased by what they heard, and felt that Apple was committed to staying competitive.
Another NDA session got into details about the internals of FullWrite. Yes, legendary FullWrite is returning from the dead! I had lunch with the principal author, who managed to wrest control of FullWrite back from Borland, where it has been bottled up for some time. It will be good to have another full-featured word processor in the Mac market.
• The Apple engineers sounded as disturbed as everyone else by the trend towards higher prices on Apple development products. They feel that the way to promote the new technologies they have painstakingly developed would be to get the needed tools into the hands of as many developers as possible, and that means cheaply and quickly. However, Apple management is much more into seeking out ways to recover costs, even if that's on the backs of the very people who are trying to make Macintosh a success.
• There were complaints about the explosion in Macintosh APIs, the technical hooks that let programmers put new goodies into their systems. In particular, AppleScript and QuickDraw GX were given almost no hope of gaining wide popularity as long as they remained staggeringly complex to implement.
• It was suggested that if anyone has had an unsatisfactory time using the 800-SOS-APPL line for hardware questions or repairs, to use the magic word invocation, "escalate". By telling the phone agent that you wish to "escalate", you will be transferred to the next higher level of supervisor.
• We should soon have a unified "PReP" standard (called PReP Plus by some) to standardize IBM-brand and Apple-brand PowerPCs. This is important because a unified standard ensures that your choice of operating systems (Mac, OS/2, Windows NT) will run on your choice of PowerPC brands (IBM, Apple and others).
One session on programming for AppleScript helped me to understand some of the thornier technical issues involved. It's sure an advantage to be able to meet the actual designers of the thing directly.
Beyond the heart of the topic, it was interesting hearing about the Windows-and-Mac integration issues involved. Prior to version 6.0 of Word, there were entirely separate development efforts behind the Windows and Mac versions. Version 6.0 brought them together into a single project, with a "common code" and "edge code" strategy. (Sorry for the buzzwords.) The speaker told of the battles to convince the PC side that the Mac users would demand certain standard behaviors, and not just sloppy techniques that have been acceptable for Windows. He also told us about the PC'ers surprise about the richness of the Mac architecture, much of which is unmatched on the Windows side. (AppleScript is a good example.)
Here's the problem: Everybody wants applications to work together, not merely to exchange data but to actually collaborate on work tasks. It's the holy grail of modern day computing. Simultaneously, programming on systems like the Mac or Windows is getting wildly complicated, with a dizzying constellation of new technologies that "just have to be supported". It's the same for users: existing applications are mighty huge these days, and we're supposed to try to get these behemoths to work together? It's sort of like breeding whales.
Key to that is that when you use your Mac or PC today, you move from program to program. Each program defines the behaviors within its own kind of documents. That is, an Excel spreadsheet is defined by Excel. In fact, it has no meaning outside of Excel. Ditto for a word processing document and its word processor program, or a media plan and a flowchart system.
Is there another way? Well, let's enter the Wayback machine, and step forward into the past (to mix metaphors from Rocky and Bullwinkle and the Firesign Theater). Jump back to the wondrously-designed but stupidly-marketed Xerox Star. The Star was even more astonishing than people realize. (Take for example its use of a touchpad to move the cursor on the screen, a dozen-and-a-half years before Apple's "breakthrough" touchpads on their newest PowerBooks.)
One of the revolutionary aspects of the Star was that it was "document centered". Rather than emphasizing programs, the Star was built around documents. Whatever you wanted to do, you started with a plain ol' empty document. You then used various tools to put things onto that document. Pick up a pen tool and draw line. Touch a text tool, and then type in some text. With a table tool, you could put in a fancy table. To the user, these tools were simple and straightforward, just tools. But to a programmer, these tools were little programs. The difference is that the document was the center of the world, and these tools were dependent on the document, not the other way around.
Return to 1994. You can copy a table from Excel to Word, and even hot link them together. But getting these massive programs to cooperate conveniently and easily is another story - it's nothing like the elegance and transparency of what Xerox did.
So the Star-like world sounds beautiful, but how do we get there? We dream of a document-based world, with many convenient tools at our disposal. But, gee, that doesn't sound remotely like the world in which Microsoft is making its billions. What Microsoft has (besides a death grip on the OS market) is a collection of hugely successful shelf-groaning applications, like Excel and Word. So Microsoft is loathe to embrace an entirely different paradigm. (Sorry to use that word, but it fits perfectly.) Thus, Microsoft gives the world (cue drum roll) OLE. OLE (which stands for Object Linking and Embedding) lets you access all the power of Excel from within your Word document, for example. Use a command to put a table in your letter, and then work with that table, but magically discovering that although you are "in" Word, you actually are using Excel. Similarly, a block of text in Excel might be processed with the full power of Word.
In this way, Microsoft gives us something that resembles the Star dream, with tools that let us "embed" functionality of all kinds into a central, paramount document. And, they can do it with existing applications, neither shocking their existing clients nor their own sales force.
The problem is that OLE is taking the elegant Star concept and shoehorning the dozen megabytes of Excel through it. We still have the ballast of great but bloated applications. And if you like Word but not its spell checker, or you like Excel but not its charting, tough cookies - it's all or nothing. Not surprisingly, OLE crawls to glacial speed on anything but the fastest PCs, and with poor facilities for network operation.
Now let's go to something completely different.
IBM, Apple, Word Perfect and others have joined together to create a separate company, CI Labs, to develop and promote another vision, called OpenDoc. In OpenDoc, documents really are generic. And tools really are concise and manageable. With OpenDoc, you take a fresh sheet from a pad of generic stationery, and use "tools" and "services" to create your work. Each of these tools is small and independent, but they work together under the common OpenDoc structure. For example, if the user changes the size of a chart on an OpenDoc document, each "part" on the document (charts, text, tables, and so on) participates in a negotiation to reformat the page.
OpenDoc builds on work from the founding companies. Its "object model" is IBM's SOM, and its strategy for interoperability is Apple's Bento, for example. It favors small developers creating insanely great tools, rather than megalithic software houses. It's completely crossplatform, so it will run on Windows, Macs and elsewhere.
And the best part is that no one vendor controls it. OpenDoc isn't crippled by having to kow-tow to the Microsofts of the world. It was said that OpenDoc fosters "an ecology of small programs" that will "shatter the role of huge applications." Strong words, perhaps a pipe dream, but a captivating one nonetheless.
At MacHack, we got a real chance to kick the tires, play with the code, and grill the creators of the OpenDoc architecture. It was exciting, and it worked. I left MacHack feeling like I just possibly had seen the future of Macs and PCs.
The Hack Show starts (by tradition) at midnight on the middle day of MacHack, and goes to wee morning hours. This year's Hack Show was interrupted at 1:30 am by the delivery of 75 Domino's pizzas, courtesy of Microsoft. This intermission itself was interrupted by some rather unfortunate behavior on the part of someone who accompanied her brother to MacHack. I'd say more, but I'd blush. In true tasteless dweeb fashion, this was recorded for posterity by a number of QuickTake electronic cameras which had been passed around at the start of the conference.
Here are some of the best entries from the Hack Show '94:
Best definitions:
Return to my Mac page.
Go to my personal page.