It was one of those gonzo ideas that I get almost every day, usually about 10:45 AM. I had been a Delphi beta tester and was itching to get a book out on it. The gonzo part of the idea was the way the book was structured: Alternating chapters, in which the first was a short project, the second containing the theory and discussion behind that project. Practice, then theory. Conventional wisdom would call that bass-ackwards, but in fact it was the way I learned Delphi: First I tried to piece something together, even if I had no idea what I was doing. Then at some point I kicked back, gathered my thoughts, and even-egad!-read the doc. It's how a lot of us learn new things, and I wanted a book that mapped onto the way I had learned Delphi.

Don Taylor's Ace Breakpoint material was just the jewel in the crown: It remains, as far as I know, the most completely insane computer tutorial ever written. Some people hated it, most people loved it. I consider it a complete success. The Delphi Programming Explorer was the best-selling book of all Coriolis Group Books titles released in the year 1995. Nothing else even came close.

I've known Jim Mischel since I moved from California to Arizona in 1990. I mentioned in my Dr. Dobb's Journal column that I was moving to Phoenix, and Jim wrote me a note welcoming me to the area. Don Taylor I've known a little longer than that, probably since 1986 or so, through the Turbo User Group (TUG) and their Get-TUG-Gether mini-conferences in Bremerton, Washington. I knew, back when I got the idea for the book, that they were both superb writers, and also men who would not flinch at the audacity and novelty of the idea.

Absolutely. I use it regularly to gen up little one-shot utilities to massage data or test out a theory. I used it to create Aardmarks, my bookmark management utility, which I use daily even though it's far from finished. I will probably be tinkering it for the rest of my life. That's good, as I've set it aside for as much as a year at a time, and yet when I go back to it, I never have to work too hard to pick up where I left off. Contrast that to my first programming language, mainframe APL, which sets into incomprehensible stone ten minutes after you lose your train of thought.

I remained with Coriolis until our investors shuttered the company in March 2002. After the magazine folded, I moved over to the books area and worked on developing the company's book product line. I trained acquisitions editors, and when the end came was designing a new killer series intended to take on both O'Reilly and Wrox. It was pretty damned slick-but it never happened. We ran out of time to save the company.

My most recent book is Jeff Duntemann's Drive-By Wi-Fi Guide, which appeared in March 2003. It's an accessible, irreverent guide to 802.11b wireless networking, with a hands-on, tinker-this-together-from-kitchen-trash slant. That's been a lot of my recent tinkering: Wi-Fi gadgetry. I've built microwave gain antennas out of those foil-lined boxes that you buy juice and chicken broth in. I've built gain antennas out of coffee cans and spaghetti sauce cans. It's all wacky good fun.

I write science fiction sporadically, and have been published for almost 30 years. I had a longish story published in Asimov's Science Fiction in April of 2002. I have an entire hard-SF high-tech adventure novel-140,000 words' worth-that's been doing the publisher rounds since mid-2000. I can't get any publishers to give it a hearing. It's not a good time to be in any sort of publishing, so first-time novelists don't have much chance. I may be reduced to publishing it myself, which I can do, but I'd really like to see it in bookstores.

As for being a contrarian, well, hey! Delphi itself is a contrarian position, in this era of ubiquitous C++. My politics (which I don't talk about much) infuriate both the left and the right, so I know I'm on the right track. The religion to which I adhere is a sort of contrarian Catholicism-we dumped the idea of a Pope in 1724, and have avoided most of the terrible troubles that have beset the Roman Catholic Church since then. (It's called "Old Catholicism" for those who may be interested.) I like sweet red wines-sweeter than your typical grim tank-car cabernet, at least. I like 60's bubble gum music. I like polyester shirts. I see what's stylish and turn hard in one direction or another. It's just the way I am.

Don threw that one at me when I least expected it. I was skeptical, but I knew Don well enough to think that he had a chance of pulling it off. And once I read the first two installments, I was convinced. It was pure brilliance in faux Seattle chiaroscuro. As for how the world saw it, well, there were few who did not have an opinion. Some sniffed and said it was stupid; most told me it was both hilarious and peculiarly effective. I've had the same reaction to some of my own work, especially the "Alien Bases" chapter in Assembly Language Step By Step, in which I create a whole new number system, complete with synthetic symbols, to teach people how to break out of their base-ten boxes. Some people can make cognitive leaps like that, and learn from unconventional pedagogy. Some can't. This won't stop me from trying new things when they occur to me.

My wife Carol and I grew up about three miles apart, me on the northern fringes of Chicago, she in a bordering suburb. We've lived all over the country: Rochester, NY; Baltimore, MD; Scotts Valley, CA; Scottsdale, AZ, and very recently (like a month ago) we moved to Colorado Springsjust in time to see TurboPower shut down. I had really been looking forward to hanging out with those guys, like I used to when we were all in Scotts Valley. If any of them read this, I hope they get in touch with me.

I took a high-school course in FORTRAN in 1969 and did very well, but it was a mainframe terminal system (obviously) and there was little I could do to pursue any skills outside the classroom. Besides, with thirty guys standing in line ahead of me, punch card stacks in hand, I confess I wasn't able to learn very much. It wasn't until early 1976 that I was able to cadge a little after-hours time on an APL terminal at the Xerox office where I worked, and the magic started to happen. APL was totally intoxicating, and I learned very quickly the pitfalls of write-only code. I wrote a 600-line text formatter utility, and by the time I got to the end of it, I had forgotten how the beginning worked. Urk. This is easy to do in APL, of course, but the lesson burned very deeply, and I've been extremely suspicious of excessively terse languages ever since. COBOL has its flaws (and I know it well) but its verbosity isn't one of them.

You wouldn't know it from my next step, though: In mid-1976 I wire-wrapped a COSMAC ELF computer on a piece of perf board, and set about writing CDP1802 binary machine code in my head. The big difference was that I documented what I was doing with flowcharts, and escaped the APL trap. I didn't even have an assembler, and by the time I had a COSMAC machine that would run BASIC, I had my first CP/M box (with a 1 MHz 8080!) on order.

And with CP/M came Pascal/MT+, a brilliant compiler that hooked me on Pascal for all time. The rest (as most of my regular readers know-I've told the story of how I came to Turbo Pascal far too often!) is history.

Professionally (that is, on salary as a titled software engineer at Xerox) five years. On the other hand, once I got my COSMAC ELF functional in 1976, there has not been a month of my life that I have not programmed something.

Pascal/MT+, BASICA, Turbo Pascal, then Borland Pascal, and for a little while, Visual Basic. Once I saw Delphi, I dropped VB like a hot rock. At this point, the Delphi rock would have to become very hot before I would drop it for anything.

In my life I've used APL, COBOL, FORTH, various flavors of BASIC, some FORTRAN, and several Xerox in-house programming languages, including Smalltalk. Z80 assembly. X86 assembly. I know enough C to hate it, and enough C++ to know I don't want to know any more. I've messed around with a multitude of others, from Java to Perl to Actor. Today I only use Delphi. I know SQL quite well, though contrary to what a lot of its groupies say, it's not a programming language. I'm exploring C#, but I won't know how well I like it until I spend some time with Sidewinder. I may use it if I really want to create .NET apps. I won't use Microsoft's compiler. They've gotten enough of my money over the years.

Well, sure. I'd also recommend careers in publishing, dental hygiene, hardware store management, and real estate sales. What I recommend to young people is to get off their asses and figure out what they love to do, and (perhaps more important) what they're good at, and chase a career in that sector. What I don't recommend is that young people go into a field just because there seems to be money in it. It's not as easy to make money in programming as it used to be, for various reasons that would make up an article in itself. Really good people who excel at programming will always find good jobs.

Back at Xerox 25 years ago, they had a program to train bored, underpaid secretaries to be COBOL programmers, and got a lot of bored, underpaid, underskilled, underinspired COBOL programmers. They got a few winners, but a lot of losers, and a lot of people were unhappy with the whole program without quite understanding why. The "why" is simple: A job is what you do to make money at any given time. A career is a passion that you can parlay into a series of satisfying, well-paying jobs. Absent that essential passion, all you have is a job, and your hold on that job is more tenuous than you think. Unless you love programming, don't do it. That goes for any other skill you could name as well.

That's a tough one. Neither programming nor computer science degree programs are anything like what they were even ten years ago, and certainly nothing like what they were when I was in college in the early 70s. What I've seen, most clearly, is that programming is not about languages at all anymore. It's about the places where huge bodies of pre-written, "black box" code come together. Hardly anybody sits down and writes a program from scratch without using loads of libraries and countless APIs. This wasn't true years ago; a good chunk of a programmer's education lay in creating his or her own general-purpose tools and libraries. Today, no matter what context you're programming in, you're mainly tying together libraries and calling operating system APIs. Every Delphi programmer knows that you can build a genuinely useful database app in Delphi without writing a single line of code. The very notion would have been mind-boggling in 1980.

So in general terms, the advice I would offer to young people seeking degrees (and careers) in programming is this: Study the edges. Figure out how operating systems and networking architectures work, and then figure out how they come together. No computer is an island anymore. It's all about edges. Even the edges have edges (look at Java's API nightmare!) and it's all a very tangled mess. Young people have to understand programming languages at a sort of higher, meta-level, so that they can pick up any arbitrary language they are required to use and get good at it quickly. You have to understand loops in the abstract so you can master them in any language.

And in truth, the mainstream, native-code languages (rather than the scripting interpreters like Perl) are converging. Basic, C, and Pascal have grown mighty damned alike in the past 15 years. Java is just C with its teeth pulled, and C# is Pascal in C's clothing. People who understand the general theory of programming languages (and by that I do not mean compiler architecture!) should be comfortable in any cabin they might choose on that particular ship.

Probably Turbo Pascal 5.5, back when I was at Borland and shortly after I left. I wrote a huge chunk of the Turbo Pascal5.5 OOP Guide, but didn't get a byline. (I got lots of money instead, and I'm fine with that!) Many people told me (once they knew I had written it) that it was the first time object-oriented concepts ever made sense to them, and I am very proud to have done that. I even named Turbo Vision. I didn't especially like it, but I named it. Prior to my suggestion it was called TOORTL. (Turbo Object-Oriented Run-Time Library.)

All the stuff I did for Xerox, as good and useful as it was in its time, is now gone and forgotten, forgotten for 20 years at the least. Remarkably, a little program I wrote in Borland Pascal and Turbo Vision, and sold for $5.95 a copy back in the early 90s, is still out there. It's called Mortgage Vision, and an amazing number of people, many of them real estate agents, still use it on creaky old 486 machines running DOS. I get email from them from time to time, asking if there's a Windows version. There isn't-and isn't likely to be, since there's no way to sell it and make the kind of money I used to make selling kioskware for DOS in 1993.

I'm working on a spam-filtering POP3 proxy (in Delphi) with Jim Mischel. Most spam filtering code glued onto mail clients these days was written by knuckleheads, and we think we can do better. Maybe we're wrong, but spam is a hideous problem. I get between 400 and 500 spams per day, and using the spam filtering in modern mail clients is agonizing. Worse, I have zero-tolerance for false positives, so I don't use Bayesian techniques or anything else shaped like a blunt instrument. We're calling it Aardmail, and we don't yet know if it will ever get released to the public. There's a lot of work to do before we'll even know if it's effective.

I don't know yet what my next book project will be. I just finished my Wi-Fi book, and I'd like to do another Delphi book, if I could think of an angle that might make money. Too soon to tell. Long on my to-do list is a "square one" tutorial on Ethernet networking. I did it for assembly language, and nobody has yet done it for networking, at least not done it to my personal standards. So that's a possibility too.

I'm still building Wi-Fi antennas from kitchen trash, and I build kites from time to time for no other reason than I just like kites.

I currently work for the new publishing company I co-founded with Keith Weiskamp, the guy I created The Coriolis Group with back in 1989. It's called Paraglyph Press. We're all Coriolis renegades, and we've got a completely virtual company. No office. Our "attic" is a directory we rent from a hosting service in Maine. We pass files around, and we make books, good books, and we make them cheaper than anybody else out there. We're doing fine, even though this may be the worst time in history to start a new publishing company. It was the contrarian thing to do.

I almost beat Bill Gates in a programming contest. Back in 1985, several high-tech editors and writers were invited up to Redmond by Microsoft's PR firm to participate in a "Storm the Gates" contest. Each of us could choose our own programming language, and Bill would use BASICA. Someone had written up several programming challenges on little quarter-sheets of paper and folded them up in a fishbowl. One of the PR women pulled a sheet from the bowl, copied it out for each of us on a copier, and we were off and running to implement the little programming problem the sheet described. It was an idiotic spec, including stuff like drawing concentric circles.

When she said "Go!" we all got to work. It was a total snap in Turbo Pascal 3.0-except that we had to read a character from the keyboard buffer "on the fly" rather than wait for it, and although I knew Turbo Pascal could do that, I also knew that it took an INT call, and I couldn't recall the details. Pascal, for the most part, doesn't work that way. So although I finished everything else first, I couldn't crack that one.

The punch line, of course, is that there's a standard function in BASICA that could do just that-in fact, there were standard functions in BASICA for every part of the challenge spec. So Gates merrily cranked along, stringing together standard functions in BASICA, and finished the spec before everybody else, though Charles Petzold got real closein text mode! (Charles faked the concentric circles by displaying various characters at the same screen position, including a degree symbol, lower-case O, and upper case O.)

It was funny in part because the specs were cooked to match the capabilities of BASICA-and doubly funny because none of us saw it coming. It was triply funny because Bill Gates obviously didn't especially enjoy the effort-clearly his PR staff had cooked up the stunt and twisted his arm to participate. He didn't stop to hang out with us much, and looked miserable throughout, even though he'd won. Now, recall that that was before Microsoft went public, and Gates was far from the world's richest man. But it was great fun to finally understand that somebody had made him do something he didn't really want to do. I wonder sometimes if that was the last time it ever happened.

Related? Being an editor at programmers' magazines. Working at PC Tech Journal was wonderful; I got to meet a lot of interesting people, including Bill Gates, Steve Ballmer, Peter Norton, Drs. Kemeny & Kurtz, and Esther Dyson. Esther was a corker. I got chewed out for challenging her at a presentation she gave, for completely misrepresenting some early work that Xerox had done in the GUI realm. Sheesh, I was there while it was happening. She had it wrong. I corrected her, and got spanked-not because I was wrong, but because Some People Are Not To Be Challenged. I'd learned that lesson before, when I was in college, and vowed I would never consort with such fools again-else I'd likely be a PhD teaching creative writing to bored frosh in a sinking liberal arts college somewhere in the butt end of Iowa by now.

Still, being an editor and a programmer at the same time was wonderful. I got to test a lot of wonderful toys for free, and my experience allowed me to understand what I was editing, the way a lot of less technical editors did not.

Trying to understand what the *&@!* clib was doing internally while I was learning to write assembly code under Linux prior to writing the current edition of Assembly Language Step By Step. Unless you're prepared to make direct kernel calls through the INT 80 call gate, you're stuck calling into clib. All the gurus say that making direct kernel calls is a no-no, and for that edition of the book, at least, I believed them. So I was forced to think like a C compiler for several months. Even using a C compiler makes me want to wash my hands with Lava soap. Thinking like oneheh, you can only imagine.

Next edition, it's INT 80. I'll explain clib calls as well, but I won't force my readers to sojourn in that particular hell if they don't choose to.

Turbo Power everything. Especially now that they've gone free and open-source, I recommend them to everyone, and am using them for everything I possibly can. There were Turbo Power tools I never had occasion to use, like Flash Filer, that I'm now learning.

I own a lot of commercial components, including Raize's DropMaster, Elevate's DBISAM, and most of what Dream Company offers, especially their data-aware treeview. These days, I'm trying to do everything possible with free components, so I can share my work with others who may not have purchased the commercial components.

Simple perseverance. Sidewinder is a good hedge against the .NET future, but I hope they continue to sell Delphi and support it. And if they do decide to shelve Delphi, I hope they take the high road and open source it, like TurboPower did with their libraries. VB would have a hard time competing against Delphi Enterprise at a price point of $0.00.

I'd be an unhappy Visual Basic programmer. Note well that I don't say "a C# programmer." C# owes its existence to Delphi. Without Delphi, C# would not have happened. I'm pretty sure of that. You can see the traces everywhere. C# is Pascal All Grown Up. I intend to learn it someday.

That's an interesting question. It would still be out there-Genius depends on little but its own brilliance-but I'm proud to say I had a (small) hand in its parentage at least, and I promoted it shamelessly in my writing and in my magazines for many years. I worked at Borland for almost two years, from early 1987 to late 1988, and I pestered them to migrate Turbo Pascal into object-oriented programming. They did, though I wasn't the only one doing the pestering.

But hey, let's get real. Delphi would still rock whether or not I had ever been born.

.NET is still pretty new and pretty green, but with Microsoft behind it, I suspect it's here to stay. C# and Sidewinder will help keep Borland in the game. We need that badly. I don't hate Microsoft, but I think that they have all the market share that they deserve (and then some) and I don't want them to get any more.

At the computer, 8 hours or more. It's how I make a living. Programming, well, it's harder to say. I sometimes pull a full, long day programming, and then don't come back to it for almost a week. I average an hour a day over the long haul, but I do it in larger chunks than that.

Newsgroups, not so much. Too noisy and spam-rotted these days, though I was an avid newsgrouper back in the mid-90's, and first had access to newsgroups back in the early 1980s when I worked at Xerox. The Web, easily two hours a day. There's no escaping it right now. It's the best way to drink from a firehose, and that's an essential skill in the 21^st Century.

Hundreds of them, in a great many different disciplines. I had to write a program in Delphi-called Aardmarks-to manage them all, in a hierarchical folder structure that runs up to seven levels deep. I'm a classification fanatic, and it's the only way I can find things when I need them. Torry, Delphi Super Page, and a multitude of API discussions, well, they're all there. I bookmark sites on other popular languages-even those I consider unnecessary or even silly-just so I can know what's going on.

There's only one way for anybody to do that: I program. I scan the Web for ideas, I watch for interesting new books, I listen to very bright guys talk about programming. I try to have a few small programming projects going at any given time. Most of them I never talk about; I lot of them I abandon long before they become useful. But being useful isn't always the point. The journey isn't always the reward, but it's still the journey-and the destination literally never stops moving.

The first several-I was at the very first one-and intermittently since. My life got complicated and difficult after 1996 for various reasons, and that's when I stopped going to most trade shows.

They kind of run together, and I can't always tell what took place when. I enjoyed meeting various Delphi powers, like Ray Konopka and Eric Harmon, at one Borland conference or another. Email makes for slightly cold contact with the rest of the human race. We all have our quirks, and I celebrate both my quirks and the quirks of others. Trade shows and conferences are one way to rise above purely textual interaction, and enjoy one another's quirks. All the Borland conferences were excellent for this, and I don't think of them as individual events so much as personal celebrations of decentralized friendships that the Net has made possible.

Michael Abrash. He's been a close friend since 1985, and I stand in awe of what he does. I've edited all of his books since 1988, so I'm very familiar with his work. I know a few other stunningly good programmers, like Anders Hejlsberg and David Stafford. Once you get up into that lofty territory, it's very hard to say what's better than what. Anders beats the rest of the world hands-down for writing compilers, but Michael doesn't do compilers. He does graphics rendering engines. David's specialty is holistic and sometimes startlingly original solutions to conventional algorithmic challenges, like file compression. Programming is increasingly a whole passel of almost completely independent disciplines. Every one has its star, and every star is very much a specialist.

Well, I'm not a world-class programmer, though I do a thorough, clean, sane job on anything I attempt. The stuff I wrote so well at Xerox was the sort of thing that would put your average champion programmer to sleep-but somebody had to write it, and I did.

My claim to fame has always been my writing, both about programming and about other things; a lot of other things, from wireless networking to ham radio to kites to the problem of woman priests in the Catholic Church. I also write SF, and have been published a fair number of times in the past thirty years, but not enough to have made a name for myself. I've been on the final Hugo Awards ballot twice for short story, but didn't win.

If I had never been a programmer, I probably would still be in publishing, though perhaps not in computer publishing. I turned down a real good job in medical publishing in 1974, right after my college graduation, because editing articles full of pictures of little old ladies with their hips laid open for surgery would have made me barf. Carol says I would have gotten used to it. I suppose-I've since gotten used to worse things, like yogurt. So I took a ten-year detour through computer technology before I got into publishing. Passion is powerful-and all roads probably lead to the same place, the place to which your passion points.

Just this:

Michael Abrash's legendary Zen of Assembly Language, which later grew into a 1500-page monster called Michael Abrash's Graphics Programming Black Book. There's nothing like it anywhere on Earth. Yes, I edited it, but that doesn't keep it from being my favorite.

Lord of the Rings. I read it every five years or so, or whenever I get sick enough to be stuck in bed for days at a time.

Hard to tell. Maybe Galaxy Quest. I'm also a big fan of Star Wars, The Matrix, The Rocketeer, and some of the recent CGI cartoons, like Shrek, Antz, and Monsters, Inc. (As for the recent Lord of the Rings movies I'm still undecided. Need to see them five or six more times!) Film for me is diversion, and nothing diverts like fantasy and SF. I don't watch a lot of "serious" film, though I did enjoy Chocolat.