Happy birthday, IBM PC!

The IBM PC 5150 turns 30 today.

IBM didn’t invent the personal computer, but if your computer has an Intel or AMD CPU in it, it’s the direct descendant of the beige box IBM unleashed on the world on August 12, 1981. Without a huge amount of effort, it’s even possible to run most of that old software on your shiny new PC. You probably wouldn’t want to, except out of curiosity, but you can do it.

I wasn’t one of the people who rushed out and got one. At the time, I was still watching Sesame Street and Mr. Rogers. I had my first experience with a computer–a Radio Shack TRS-80–in 1982, and the first computer my family bought was a Commodore 64 in 1984. Even in 1984, there were still plenty of people who questioned why anyone needed a computer in their home. My introduction to the IBM PC and PC-DOS didn’t happen until 1987.

The IBM PC, in its standard early 1980s configuration
IBM built its original IBM PC, introduced in 1981, using mostly off-the-shelf parts.

That original PC did a lot of things wrong and a lot of things right. The power supply was a really wimpy 63 watts, and it only had five expansion slots, which really limited its capabilities until so-called multifunction cards, which combined things like serial ports, parallel ports, memory, and sometimes even the disk controller onto a single card to save slots. Even then, with only five slots to work with, you couldn’t really turn that original PC into a dream machine. You couldn’t add a hard disk to it without getting a kludgy expansion chassis that was just as big as the 5150 itself. But in IBM’s defense, they designed it with a five-year life expectancy in mind and only anticipated selling 250,000 units per year.

It ended up doing a lot better than that.

IBM legitimized the market. There were standards before IBM came along, of course. Most business machines ran an operating system called CP/M, and most software that ran on one CP/M machine would run on the rest of them. But all of them had slightly different disk formats and other nagging inconsistencies.

Over the course of the decade, IBM made those sorts of inconsistencies unacceptable. Similar machines appeared very quickly, but people weren’t willing to tolerate another ecosystem like those early CP/M machines. People wanted 100% IBM compatibility, or very close to it, and eventually they got it. Compaq appeared in 1982 with the first true IBM-compatible computer and undercut the prices slightly. But even by the mid 1980s, you had to be pretty wealthy to afford even an IBM clone, let alone a true-blue IBM. I knew families who had them, but we’re talking the families of surgeons, trial lawyers, and the owner of the town’s most popular restaurant and rental place.

Most of the people I knew were perfectly happy with 8-bit Commodores.

But the future belonged to IBM and the cloners. And it’s easy to criticize IBM for making the PC too easy to clone, but easy is a relative term. In 1986, fully equipped clones that cost less than $1,500 started to appear. The Leading Edge Model D was one of the most popular of those clones, and it received tremendous reviews. But within a matter of weeks, hordes of people appeared waving software titles that didn’t run properly on the Model D.

The maker of the BIOS–the key to IBM compatibility–came up with a revision to improve compatibility, and software developers started testing their software on the Model D in addition to the IBM PC. They recognized that the sub-$1,500 clones would soon outsell the original. If you wanted to sell a lot of software, testing only on the original soon wasn’t enough. That led to some growing pains, and an IBM engineer once telling me, “Even IBM computers aren’t 100% IBM compatible anymore.” Eventually, software solved most of that problem.

And, frankly, to claim IBM would have been better off without the clones is to ignore history. No other computer architecture from the 1980s survived. The Macintosh still exists, but the Macintosh of 1984 bears little resemblance to the Macintosh of 2011. You can’t run software from that machine on a new Macintosh. Getting DOS 1.0 to run on the motherboard I bought last week would take some doing, but I could run DOS 4 on it (and perhaps DOS 3.3) and the original Visicalc and other apps from 1981 will run once I get DOS up and running.

The IBM PC certainly exceeded expectations, lasting past 1986 and 1.25 million units sold.

Looking back now, I think whatever standard had been the most open would have eventually won. IBM happened to have the most open architecture, and that’s what won. And where IBM misstepped was when they refused in 1985-1986 to create a state-of-the-art PC based on Intel’s 32-bit 80386 CPU. Compaq did, and became the visionary, high-tech leader. Then in 1987, IBM released the Personal System/2 (PS/2), which ran all of the same software as the IBM PC, but introduced a new, proprietary, clone-proof architecture. Again, Compaq stepped in, and teamed up with some others to develop an open standard that was nearly as good, backward compatible, and cheaper. IBM lost a lot of goodwill with that move, and once they reversed course, it was too late. In the hearts and minds of the buying public, IBM was proprietary and expensive while clones were neither.

If IBM had taken a more PS/2-like approach in 1981, it would have been just another quaint machine from the 1980s.

The IBM PC wasn’t a significant improvement over what came before it. It took the best elements of the day’s technology–the Apple II’s expansion slots, CP/M’s BIOS–and was, pretty much, good enough. Whether by design or accident, it proved extensible enough to remain good enough as times changed. Ultimately that’s why it succeeded. Better things came along as the decade wore on, but the IBM PC’s design always allowed them to bolt on some addition that was good enough to compete.

There’s a lot more to the IBM PC story, including how Microsoft came to supply the operating system. I’ve explored that topic before, and longtime PC Magazine editor Michael Miller presents a reasonable, balanced account. I have heard rumors over the decades that PC DOS 1.0 contained an easter egg that contemporary versions of CP/M also contained, which would completely vindicate Gary Kildall’s allegations, but nobody has ever come forward with details. The details didn’t come out on the PC’s 25th anniversary, and if they don’t come out this week, one pretty much has to assume it’s just a rumor. All we have is Gary Kildall’s quote–“ask [insert Microsoft employee name here] why function code 6 ends in a dollar sign. Nobody in the world knows that but me.”

Even given that widely quoted hint, I can’t even find any thorough analysis or comparison of function code 6 online. Kildall died in 1994, so we won’t be hearing any more from him about that.

But back to the product. IBM never recovered from the lost momentum in the late 1980s due to the 386 and the PS/2. By the time I started my professional career in the mid 1990s, the majority of computers IBM sold were no different from anyone else’s PCs.They used the same parts, and even the price was pretty close. But they weren’t making a whole lot of money doing it. I wondered sometimes in the 1990s why IBM didn’t try to leverage its vertical integration more. IBM was making CPUs for Cyrix and got half of each batch out of the deal, but IBM sold them on the open market rather than putting them in its PCs. IBM made hard drives, but when you opened up a 1990s IBM PC, you were just as likely to see a Quantum or a Maxtor hard drive in there as an IBM drive. Maybe they were afraid that if they put too much of their own stuff in their PCs, they’d look too proprietary.

IBM hung on for a while, justifying the money it was losing by saying that selling PCs helped them sell services. But in 2004, IBM gave up and sold its PC division to Lenovo.

I guess a lot of people see that as a defeat, but look how few companies that sold computers in the 1980s remain. I count three: Apple, Dell, and HP. The rest exist only as brand names for some other conglomerate. Computers aren’t very profitable for anyone else anymore either, though some of the parts that go in them remain immensely profitable, as are some of the things that plug into them and the software they run.

IBM certainly took some missteps, and has been a non-factor in the PC business for long enough that people of a certain age might not understand the phrase “IBM compatible.” But there’s no doubt, even though it later lost control of it, that 30 years ago, when IBM launched its PC, it created a standard and it changed the industry.

%d bloggers like this:
WordPress Appliance - Powered by TurnKey Linux