Troubleshooting Professional Magazine
|
Where Have All the Heroes Gone? |
When the magazine hit the newsstand, orders came rolling in, $400.00 each order. By February there were 250 orders per day. The computer revolution had begun, and the race was on. The Homebrew Computer Club quickly formed in Silicon Valley. Bill Gates and Paul Allen dropped what they were doing and moved down to Albuquerque to be part of the action. Others began planning their own personal computers.
One could say that Ed Roberts led the computer revolution, but that's not exactly it. If he hadn't done it, someone else soon would have. They were lined up deep behind him. What Ed did was more like firing the starters pistol, and then participate in the race. MITS held out for two years before being trampled by the competitive hordes. His competitors took his 100 line Altair Bus, promoted it as the "S100 Bus", and steamrolled him, using Gary Kildall's CP/M. MITS never saw the 1980's.
1975 is long gone, and Ed Roberts has a new life today. He's a physician. But I wonder if when he's alone, thinking of computers, just for fun he holds up his right hand and utters the words "racers, take your mark".
And that puppy cooked! The greased-lightning Wordstar word processor stayed ahead of every keystroke, paste and format. Turbo Pascal 2.0 that compiled and ran faster than the DEC PDP 11-23's I used at work. And the sweetest little operating system you could ever wish for.
It was called CP/M, sold by Digital Research. Everything was built for efficiency. CP/M had this ultra-efficient PIP program whose business in life was copying data from here to there. As I remember it, it resembled modern day Linux' dd command. That little suitcase machine seemed quicker and nimbler than a minicomputer. And it was all mine.
Digital Research was headed by Gary Kildall. A Seattle native, in 1973 he wrote PL/M, (Programming Language for Microcomputers), and then wrote CP/M (Control Program for Microcomputers) in PL/M so he'd have the services needed by PL/M. Of course, C and Unix had formalized the chicken and egg development model years before, so he was on firm ground.
In or around 1974 he began actually selling CP/M on floppies, and by 1976 had founded Digital Research. Remember, those were the days when a "personal" computer was more often than not a kit, often using an audio tape recorder or something similarly hoaky as storage. And if you wanted software you wrote it yourself. In assembler. CP/M changed all that by supporting floppies and hard disks. It's true the OS came with an assembler and if you wanted to add a new device you needed to patch the OS, but CP/M was brilliantly portable and ahead of its time.
By the late 70's 8080 and Z80 based computers were selling like hotcakes, primarily because CP/M made them usable. Prices were coming down, and the average Joe could buy one.
Then there was "the IBM thing". In mid-1980 IBM approached Digital Research about using CP/M as the OS for the planned "IBM PC". It didn't happen. Everyone knows part of the story, the only guy who knows the whole thing is gone. IBM went on to use "MS-DOS", version 1 of which bore an amazing resemblance to Seattle Computer Products' QDOS. But in the early 80's, Gary Kildall's OS ruled the world. Kaypro, Morrow Designs, Osbourne, and a host of other clones sole amazingly cheap, amazingly capable CP/M machines. A generation of budding programmers learned on these machines.
The party lasted til 1985 or 1986, when it became obvious MS-DOS controlled the OS world. Digital Research came out with a task switching DOS clone, called DR-DOS, in the late 80's. I used it, and I can tell you it was wonderful. When Windows 3.0 kidnapped the world, I even took the extra time and trouble to get Windows to install on top of DR-DOS (and wasn't that an interesting install). Most computer users couldn't or wouldn't do that install, so DR-DOS became Windows' first victim.
Gary Kildall died in July of 1994, at the age of 52. His world, the world of Kaypro, Morrow Designs, Osbourne and CP/M are gone. But an entire generation of technologists remember, and we remember who REALLY did the innovation.
The Apple II, a real computer you could use without a soldering iron, came out in 1977, soon after Apple incorporated and moved their headquarters out of Jobs' garage. By mid 1978 you could buy a floppy for it -- the days of tape and toggle were over. In 1979, Personal Software releases the first "Killer App", the VisiCalc spreadsheet, for the Apple II. It was written by software pioneers Dan Bricklin and Bob Frankston. The question "but what do I do with it?" was finally answered. In July of 1981 Dick Cavett became the spokesman for Apple. This would be like having Jay Leno for a spokesman today. Apple would eventually become the number two computer vendor, behind IBM.
Steve Wozniak resigned from Apple in February of 1985. Steve Jobs left in September of that year. Apple slogged along for years. In 1996 a desparately sick Apple hired Jobs as a consultant, and hired him as CEO in 1997. Jobs resuscitated Apple back into the black. Apple recently sold 800,000 iMacs in six months.
And Wozniak? He's a fifth grade school teacher. Boy, I wish I'd had a fifth grade teacher like that.
The next few years are spent marketing Basic, which becomes the de-facto standard microcomputer language. In the late 70's, if you programmed on any micro besides an apple, you were probably using a product from Paul and Bill. With the demise of MITS, there was no reason to stay in Albuquerque. Microsoft moved to Washington state in early 1979.
The IBM-PC came out in 1981, sporting the MS-DOS operating system, which bore a striking resemblance to QDOS from Seattle Computer, which itself bore a remarkable resemblance to CP/M. There are various stories on how this came about. The important thing about MS-DOS was the config.sys file, into which references to stand-alone drivers could be placed. For the first time in the Intel/Zilog world, a hardware change could be done without an assembly language tweak by the user. Finally, upgradable computers were available to the masses.
MS-DOS created a single target platform for the independent programmer. At a time when a Unix machine could cost tens of thousands, a fantastic DOS machine could be had for $2500. DOS created the programmer cottage industry and significantly raised the skill bar. No longer would industry stand still for the stodgy mainframe development cycle.
In 1990 Microsoft introduced Windows 3.0. It had multitasking, easy to use cut and paste, and fonts built into the operating system. It looked like Microsoft owned the world. But safely off Microsoft's radar, Richard, Linus, Marc, Larry, Guido, and an army of others quietly worked on their own plans.
Dan Bricklin and Bob Frankston designed and coded VisiCalc in 1978 and early 1979 under contract, and it was released in May of 1979. VisiCalc turned out to be the KILLER APP, justifying business decisions to buy computers. These two guys changed the world forever. Computers were no longer a hobby.
Unfortunately, VisiCorp, VisiCalc's publisher, sued Software Arts, i.e. Dan and Bob, in 1983. There was a countersuit, and everyone lost except the lawyers and Lotus Development Corporation, to whom Software Arts' assets were sold. Bob went to work for Lotus, Dan went on to create the Dan Bricklin Demo program. VisiCalc became the excuse everyone needed to buy a computer, spurring the early 80's boom. The rest is history.
Except they're at it again. Dan Bricklin has a web content creation system called Trellix, which makes nice looking, informative and snappy web pages. And Bob Frankston, after stints with Lotus and Microsoft, is championing an improved IP specification and educating the populace on Internet issues.
In those days I was an apprentice programmer. I'd work all day on a PDP 11, using this [name deleted] Pascal that would, using the same piece of code, compile 7 out of 10 times (you guess which times). To compile the whole system took a half hour if, by chance, the compiler didn't crash.
At night I'd go home and do Turbo Pascal on my Kaypro 2x, compiling in 10 seconds on floppies with a 2mhz processor and 64K of ram. Two years later I used Turbo Pascal to write the front end of what would have been have been called a client-server app, if the term had been invented yet.
Turbo C came next, then Turbo C++. I learned my first OOP from a video where Philippe discusses objects armed with a Corvette and his clarinet. Turbo C++ was my language of choice when I wrote my first OOP system in 1991. It handled two hundred million dollar annual revenue until 1997, when the client switched accounting systems and needed it written again. Did I switch to Microsoft Visual C++ with MFC? Not a chance. With that kind of money flowing through the system, reliability is a premium. Turbo C++ again!
Philippe Kahn resigned as president and CEO of Borland in January, 1995. Since then the company changed their name, forgot who their customers were, and generally floundered. Apple, when faced with an almost identical situation, brought back their founder to turn things around. Inprise recently found themselves without a CEO. If they have a lick of sense, they'll bring back Philippe, and the name he made famous.
Kind of reminds me what 1999 would be like if Richard Stallman hadn't written the GNU Manifesto in 1984. It advocated free software, specifically a free UNIX workalike. In the Manifesto, Stallman prophetically described the process of getting this to happen, including ideas in licensing (must pass on source and all rights to the receiver, etc). In 1991 he copyrighted the GNU General Public License (otherwise known as GPL). That license provided a framework allowing a developer to guarantee that his work would never be co-opted or subverted by an unscrupulous corporation. Software authors began to license their software using GPL.
Can you imagine life today if he hadn't done this? Like Linux? Thank Stallman. Like Python, Perl, TCL or Apache? Thank Stallman. Without him this would be a truly ugly world.
OK, Mr. Ninja manager, try this. I'll give you 8 years to develop the world's best operating system, make it commercially viable, and convince major players to write software for it and bundle it with their machines. You can hire as many programmers as you want, but your budget will be zero. Nothing. Nada.
Can't do it? The techie who wrote the Linux kernel did it. He managed the activities of thousands of programmers to bring a cohesive, commercially viable product to the world, possibly upsetting the strongest technology monopoly in history.
How did Linus Torvalds do it? In his famous "The Cathedral and the Bazaar" paper, Eric S. Raymond explains it better than I could. The URL is at the bottom of the page.
Richard Stallman wrote the manifesto. Linus Torvalds proved it worked.
| Perl | Python | |
| Development speed | 186,000 miles per second | 186,000 miles per second |
| Reliability | Almost perfect | Almost perfect |
| Regular expressions | Complete | Complete |
| Object orientation | Excellent, just short of Python | Excellent, just short of C++ and Java |
| Database connectivity | Ubiquitious (via DBI::DBD) | Catching up fast |
| Price | $0.00 | $0.00 |
| Support | Outstanding | Outstanding |
| Bundling | Ubiquitous in Linux | Ubiquitous in Linux |
| Portability | Linux, Unix, Windows, Mac | Linux, Unix, Windows, Mac |
| XML Support | Yes | Yes |
| CGI scripting | Yes | Yes |
I'm continually asked which should be used. Here are my guidelines:
Larry and Guido made it possible.
Right around that time a kid named Marc Andreessen left home to attend the University of Illinois at Champaign. In 1992 he got a $6.00 an hour job. But not just any $6.00/hr job. Nope, Marc worked at the National Center for Supercomputing Applications, where he started working on this app to make an easily authorable visual interface to TCP/IP networks. The interface was called Mosaic, the first web browser. It was released to the public just before Marc graduated college.
In early 1994 Marc and Silicon Graphics founder Jim Clark formed Netscape and released Netscape Navigator. Within a year their browser spurred the first big internet wave. Three years later most middle class households were online.
So that's how this kid in his twenties darned near knocked off Microsoft. Microsoft survived the Netscape explosion only by throwing much of its huge wealth at quickly creating a competing browser, and waging an economic war of attrition on Netscape. So Netscape went free-software and still survives -- the darling of free software.
I got in a time bind and needed to write this issue of Troubleshooting Professional in a single day. Do you think I would have stood a chance without my handy dandy Netscape web browser? Here's another question. If Andreeson hadn't created Mosaic and co-founded Netscape, do you think Internet Explorer would exist in a usable form today? Would you be reading this article?
Did you know that a heavy duty data enabled web development environment, called Zope, has surfaced as free software within the last year? Did you know that a group is now writing a portable, free software Clipper workalike? Did you know that both Perl and Python have XML interfaces already? Do you think five years from now some of these people might be considered heros?
So next time time you're at your Linux User Group meeting, look around the room. That guy sitting next to you just might be the next hero. Or maybe the next hero is even closer. Think about that the next time you look in the mirror.
All submissions become the property of the publisher (Steve Litt), unless other arrangements are previously made in writing. We do not currently pay for articles. Troubleshooters.Com reserves the right to edit any submission for clarity or brevity. Any published article will include a two sentence description of the author, a hypertext link to his or her email, and a phone number if desired. Upon request, we will include a hypertext link, at the end of the magazine issue, to the author's website, providing that website meets the Troubleshooters.Com criteria for links and that the author's website first links to Troubleshooters.Com.
Submissions should be emailed to Steve Litt's email address, with subject line Article Submission. The first paragraph of your message should read as follows (unless other arrangements are previously made in writing):