Dr. Dobb's Journal January 2001
Data n pl: original Latin meaning: something given.
Base adj: of low or no value; worthless.
Database n: worthless gift (see above).
This month and next, I offer a mixed bag containing snippets of information about database-design products, business rules, new programming environments, and (this month only?) the cottage industry of writing books about Steve Jobs. After Apple's recent stock slide, it looks like it's already time for another one of those.
By now, either you have read Alan Deutschman's The Second Coming of Steve Jobs (Broadway Books/Random House, 2000; ISBN 0-7679-0432-X) — or you have no intention of doing so. If the former case applies, you may find something of value in the following exegesis of the book in the context of other hatchet jobs, love letters, and keen analyses on the subject of SPJ. If the latter case applies, this may give you something to talk about when other people ask you what you thought of the book.
Sales of Deutschman's book were doubtless helped by the reports that Steve Jobs had twisted arms to keep Vanity Fair from running an excerpt of the book. A catchy and evocative phrase is a good marketing tool, too, so whoever came up with the phrase "Sharon Stone incident" to describe one sordid event reported in the book probably didn't hurt sales any, either. But it's substance, not public relations, that matters in the long run, which is why Ralph Nader is President of the United States.
Excuse me, I think I'm starting out with entirely the wrong attitude here.
Deutschman starts with Jobs's departure from Apple 15 years ago, tells the stories of the founding and growth of NeXT and Pixar, follows Steve back to Apple and Apple back to profitability, and ends with some not altogether consistent reflections on what makes Steve Steve.
Actually, though, the attempts to explain Steve Jobs pervade the book, as they do most books about Jobs or about Apple. Not that there's anything wrong with that.
I don't have any major quarrel with the book, factually or artistically. Okay, there are a few grammatical errors; things like "lay" for "lie," "rode" for "ridden." And technically, the Pixar chapter is flawed because it totally abandons the book's subject, Steve Jobs, for 20 pages while it recounts the entire history of Pixar before Jobs came on the scene. But that story is so fascinating, it would be churlish to complain.
Not as churlish as brow-beating the editor of Vanity Fair over that excerpt, but churlish.
On balance, is Deutschman fair to Steve Jobs? I don't know if that's an achievable goal. Jobs has his supporters and his detractors, and nobody in either of those groups does justice to his complexity, to his peculiar mix of strengths and weaknesses. Deutschman tries to present both sides, and does better than most writers in maintaining a balanced view. Despite this, he rips Jobs open like the hole in the Hindenburg, to use John Dvorak's metaphor.
My own take on Jobs would emphasize the positive a bit more. Steve Jobs himself was the one essential element in Apple's recent recovery, he was the defining and driving force behind the Macintosh, he was the entrepreneur who created Apple Computer. He empowered fine work at NeXT even if the business was an utter failure. He supported breakthrough work at Pixar and turned it into a Hollywood player through his negotiating powers. Lee Butcher's 1987 book on Jobs was called Accidental Millionaire. It's now 14 years and a lot of happy accidents later. When the accidents keep happening, you have to acknowledge that they aren't really accidents.
On the other side of the ledger, Jobs can be a world-class jerk. It's sort of a half-full/half-empty thing, isn't it?
The hardest part about writing a book on Steve Jobs, apparently, is the last chapter. Nobody has yet written a successful last chapter on Steve — or on Apple — although a lot of people have tried. For example:
It was clear [in 1984] that Apple would fight to maintain its technological edge, or at least the very illusion of a technological edge...Jobs liked to say that the computers Apple would make in five years would look nothing like the machines they were selling...
— Michael Moritz, The Little Kingdom, 1984
Jobs may have liked to say that, but the machines that Apple was making five years later in 1989 were unremarkable later versions of the same old Macintosh, just as they are today. Of course, Jobs had nothing to say about Apple products during those five years, having been forced out of power and out of the company in 1985, which Moritz didn't foresee. As for Apple fighting to maintain its technological edge, the company soon locked into a technological clinch, while Microsoft kept jabbing away until Windows was Maclike enough to destroy any illusion of a technological edge...
At age 30, Jobs had been deposed in a bitter power struggle...Apple, on the other hand, was looking forward to a much brighter future without him.
— Lee Butcher, Accidental Millionaire, 1987
You have to do the numbers awfully carefully to support that "much brighter" prediction. At the release of the Mac in 1984, Apple looked like it was in the catbird seat; and 10 years later, it looked like something scraped off the floor of a birdhouse. At the time Butcher was writing, though, Apple had recently been in trouble, things did indeed improve when he left, and it's pretty clear that this was cause and effect. Give Butcher that one. But that wasn't his final chapter. In his final chapter, he says:
Apple started 1987 as a two billion dollar company that had John Sculley's personal mark on every nook and cranny...The casual dress belies the discipline that Sculley demands, but the discipline does not seem to have dampened creativity.
— Lee Butcher, Accidental Millionaire, 1987
In fact, this supposed discipline never took root at Apple, which is why the company was pathologically unable to deliver a new operating system. As for Sculley's personal mark, the most obvious evidence of that was his pet technofantasy, Knowledge Navigator, which soon started sucking up money in the form of the Newton project:
[C]onsidering Apple's disarray at Newton's introduction in August 1993 — Sculley no longer at the helm, stock price down, layoffs in the air — it seemed that Apple's future indeed rested with Newton and with other instances of what it called "personal digital assistants."
— Steven Levy, Insanely Great, 1994
If Newton had saved Apple in 1994, it would have fulfilled Jobs's prediction that the "computers Apple would make in five years would look nothing like the machines they were selling [in 1984]" half a decade late. But as hindsight so plainly tells us, Newton was never going to pull Apple's chestnuts out of the fire, and had died of premature release and PR poisoning long before the plug was pulled on it in 1998 by none other than interim CEO Steve Jobs:
By the time this book is published, the question of who will take over [as permanent CEO] should have been answered ...but here are some of my possibilities: Bob Frankenberg...Eric Schmidt...Del Yocam..."
The Wintel market forces are just too overwhelming, I believe, for Apple not to be forced into a merger with a stronger partner. It could eke out a living for a few more years on its own...
— Jim Carlton, Apple, 1997
By the time Carlton's book was published the question was indeed answered, and Carlton was snookered like the rest of us by Steve Jobs's coy dance in which he slowly stripped the "i" off his title, getting all the publicity possible out of turning from interim CEO with all the power into CEO with exactly the same power. As for the Wintel market forces squeezing Apple into a merger, it could still happen, but the recent spectacular turnaround at Apple was nowhere on Carlton's radar when he wrote this:
Steve had...sold two million iMacs in only twelve months. And he had quintupled Apple's stock price in two years...Steve Jobs had finally achieved his vindication."
— Alan Deutschman, The Second Coming of Steve Jobs, 2000
Oops, there's that dangerous word "finally" again. Soon after Deutschman's book came out, Apple announced that it would not meet quarterly profit expectations and half that stock price increase went away overnight, cutting the legs out from under any "vindication," and raising all the old questions about Apple's long-term viability and Steve Jobs's managerial skills.
Okay, this is fun, but I have to admit that it's way easy to make people look foolish when they attempt to predict the future. Especially regarding a technology company. Especially when that technology company is Apple. May I recommend my own favorite Apple book? It's Owen Linzmyer's Apple Confidential (No Starch Press, 1999). Owen makes only one prediction, so far as I can tell, in the entire book, and it's in the last sentence of his last chapter:
[I]f the past is any guide, you can be sure that Apple's future will be anything but dull.
— Owen Linzmayer, Apple Confidential, 1999
In a recent column, I wrote about databases, business rules, and Prolog. The column elicited some knowledgeable feedback from readers. These are not worthless gifts.
Michael Covington wrote to tell me that Prolog is very much alive. The reason you don't see a lot of advertisements for Prolog compilers, he said, is the high quality of the available free ones. SWI is well-established; GNU up-and-coming; LPA and SICStus are the leading commercial Prolog implementators. At http://www.ai.uga.edu/~mc/wizard.html, Covington demonstrates how to put natural language processing into a CGI script with Prolog, which, along with implementing business rules, is another "natural" application for Prolog.
Andrew McKinlay, the chief architect of Suneido, thought that given my interest in databases and programming, I might want to take a look at his baby. Suneido (http://www.suneido.com/) is an open source (GPL) object-oriented language plus client-server relational database plus application framework, written in C++ for Windows 95/98/ME/NT/2000. (The server requires NT or 2000.) As McKinlay describes it, Suneido was designed specifically for developing business applications and information systems. The integration of database and development system in Suneido is particularly useful when implementing business rules. "The whole point of having Business Rules," McKinlay says, "is to have that logic in one place, usable by the entire system. Unless the system is completely integrated, you don't really get the true benefits of that." The Suneido IDE was written in Suneido, so since it's all GPL, you get the full source for that.
I also heard from the Rebol horde. Rebol Technologies (http://www.rebol.com/) released Rebol/Command in October. The original product, Rebol/Core, remains free and available on a wide range of platforms. Rebol/Command and Rebol/Command Runtime are commercial products that offer access to platform-specific libraries and tools and third-party apps. "Platform-specific libraries" is a tip-off that this version isn't platform neutral like Rebol/ Core. It's currently available for Windows, Linux, and Solaris. The Windows version gives direct access to relational databases via ODBC calls. The UNIX versions talk directly to Oracle.
The latest missive from the folks at FileMaker also emphasizes compatibility with the Borg, I mean Microsoft. FileMaker Pro is a cross-platform database system, but it started out on the Mac. The Mac implementation has been getting more of a Microsoft look, which doesn't please all users. The latest news has to do with closer integration with Office for the Mac, which is surely a good thing. With the latest versions of both products, you can now just drag a FileMaker Pro file onto Excel, and a Wizard walks you through importing records into an Excel spreadsheet; and you can later refresh the data in Excel if it's changed in the database. You can also get FileMaker data into Excel via the Open command in the File menu or the Get External Data command in the Data menu.
As for the possibility of a UNIX version of FileMaker Pro, those who would know aren't talking, but current evidence says that they are much more interested in the Palm market. By the time you read this, FileMaker may already have shipped FileMaker Mobile, which lets you access databases from Palm devices. One UNIX variant will definitely get a FileMaker port, though: FileMaker plans to ship a release for MacOS X as soon as possible after OS X ships. This will, I believe, be a Carbonized version, like most of the early OS X apps. We've been using FileMaker here at The Prose Garden for a while now, and for quick-and-dirty, get-the-job-done-today database development, I can't imagine a better tool. It's also a much better relational database development system than it was. And if you want to put a database on the Web, as we do, FileMaker makes the job downright easy.
That's the upside; here's the downside: FileMaker took a lot of heat a year ago when it introduced Version 5. As Version 5, FileMaker fragmented into four different products: FMP 5, FMP 5 Server, FMP 5 Unlimited, and FMP 5 Developer. The base product costs $250, Developer $500, and the other two are a thousand dollars each. In splitting up the line, FileMaker was attempting to herd users into different pens.
Running a small operation with 10 or fewer users, either on an intranet or in a small workgroup LAN? Never need more than 50 files open simultaneously? You want FMP 5. Want to publish on the Web or to an unlimited number of intranet users? You need FMP 5 Unlimited. Concerned about performance? Need to serve a lot of network users and/or have a lot of files open simultaneously? Want to be able to administer the database remotely? You need FMP 5 Server. Want to build custom web solutions with XML, JDBC, ODBC? Want to create custom front-ends to your databases with RAD tools? Distribute royalty-free run times? You need FMP 5 Developer.
Those are, in my opinion, reasonable product categories, and for many users, these categories made perfect sense. But for some, the change was an outrage. One inexpensive version of FMP 4 had been enough for all users in the past. The new product and pricing plan was met with a blizzard of cost scenarios angrily posted to various newsgroups and web sites, showing how much it would cost small businesses to upgrade to Version 5. Certainly, if you had 11 users or had planned to use one copy of FileMaker 4 to publish a database on the web, upgrading to Version 5 was going to be hard to justify. And FileMaker had changed the file format with Version 5, so FMP 4 would not be able to read databases created in FMP 5. In the past year, FileMaker has sold a lot of copies of its various FMP 5 products, but there's a contingent of diehards out there who are sticking with V4.
We planned to put our database on the Web using an ISP, so the choice of a FileMaker version was out of our hands. Our ISP made it clear that he wasn't interested in moving up to the expensive FMP 5 Unlimited. FMP 4 worked fine for him, he said.
We were already developing in FMP 5 and using it on our less-than-10-user LAN, and given that we were willing to update the database only once a day, it was not totally crazy to imagine moving the data back and forth between FMP 5 and FMP 4 databases. Besides simply ex/importing the data, we had the option of using FMP Reverter, a third-party tool that provides backward compatibility from FMP 5 to FMP 4.
But it seemed odd that, with FileMaker's efforts to integrate FileMaker so tightly with Office apps, we had to treat a different version of FileMaker as though it were some alien application, and devise tricks for connecting with it. Was FMP 5 a worthless gift? We didn't think so, but we definitely had some issues to address.
That's the position we'd gotten ourselves into by the time I needed to turn in this column. I'll report on chapter two of this saga next month.
DDJ