Monday, December 18, 2006
"Innovation" at MSFT
A central feature of MSFT's corporate imagemaking is "innovation." MSFT has to charge a lot to support its "innovation." MSFT can't be regulated or obey antitrust laws because they stifle "innovation." The "free market" encourages MSFT to "innovate." And they've got a huge stack of patents which "prove" how much "innovating" is happening at MSFT all the time.
But what of value has MSFT actually invented? Their main operating system product is essentially a bad imitation of unix, and the places where it differs from unix are generally regarded as poor design decisions. The desktop "metaphor" is from Project Athena at MIT, via Xerox and Apple and a bunch of engineering workstation companies you never heard of. Distributed Common Objects (DCOM and .NET and much of Office) is from the Xerox Alto/Star/Expert networked personal workstation. The font rendering stuff goes way back. Sun had dynamically linked shared libraries long before MSFT, and did a better job, and Linux had it in production a year before the first Windows-95 beta. Package management with dependency control? MSFT still doesn't have it. Multi user security, enforced by the kernel? Unix had that in 1970, and wasn't first. Maybe it'll be usable in Vista, but don't hold your breath.
It reminds me of the company that became General Electric, when Edison was running it. They avoided original ideas, because of the risk. Instead, they specialized in "bringing" other people's ideas "to market." That is, cheating the original inventors. Consider the light bulb. Incandescent bulbs were used in high end stage lighting and store window displays before Edison opened its doors. They were made by hand and cost too much to compete economically with gas and arc lights. What Edison did was hire engineers on salary to develop a light bulb design that could be manufactured by a machine. Then he put on a huge public relations campaign to make people think he'd personally invented the incandescent light bulb. Then they sold them below manufacturing cost to create demand for electric power and distribution equipment. Same with gramophone cylinders and the telephone.
Skimming through some of the patent abstracts folks are posting in the debate over software patents in Europe, I see basically two kinds. There are those with plenty of prior art, which, I suppose, MSFT just expects to get away with because they have more lawyers than anybody else. Those which are obvious. Things that are so obscure there's no obvious value. That kind aren't patentable unless you bend the original rules for patents past the breaking point. Must be novel, original, and valuable. The second kind all have to do with copy protection, thwarting reverse engineering, and generally enforcing copyright and other intellectual property monopolies. I'll give MSFT credit for "innovating" in that field. Those inventions are of value to monopolists, but it's hard to see the value to the public. The one thing Edison is known to have personally invented was the stock ticker.
But what of value has MSFT actually invented? Their main operating system product is essentially a bad imitation of unix, and the places where it differs from unix are generally regarded as poor design decisions. The desktop "metaphor" is from Project Athena at MIT, via Xerox and Apple and a bunch of engineering workstation companies you never heard of. Distributed Common Objects (DCOM and .NET and much of Office) is from the Xerox Alto/Star/Expert networked personal workstation. The font rendering stuff goes way back. Sun had dynamically linked shared libraries long before MSFT, and did a better job, and Linux had it in production a year before the first Windows-95 beta. Package management with dependency control? MSFT still doesn't have it. Multi user security, enforced by the kernel? Unix had that in 1970, and wasn't first. Maybe it'll be usable in Vista, but don't hold your breath.
It reminds me of the company that became General Electric, when Edison was running it. They avoided original ideas, because of the risk. Instead, they specialized in "bringing" other people's ideas "to market." That is, cheating the original inventors. Consider the light bulb. Incandescent bulbs were used in high end stage lighting and store window displays before Edison opened its doors. They were made by hand and cost too much to compete economically with gas and arc lights. What Edison did was hire engineers on salary to develop a light bulb design that could be manufactured by a machine. Then he put on a huge public relations campaign to make people think he'd personally invented the incandescent light bulb. Then they sold them below manufacturing cost to create demand for electric power and distribution equipment. Same with gramophone cylinders and the telephone.
Skimming through some of the patent abstracts folks are posting in the debate over software patents in Europe, I see basically two kinds. There are those with plenty of prior art, which, I suppose, MSFT just expects to get away with because they have more lawyers than anybody else. Those which are obvious. Things that are so obscure there's no obvious value. That kind aren't patentable unless you bend the original rules for patents past the breaking point. Must be novel, original, and valuable. The second kind all have to do with copy protection, thwarting reverse engineering, and generally enforcing copyright and other intellectual property monopolies. I'll give MSFT credit for "innovating" in that field. Those inventions are of value to monopolists, but it's hard to see the value to the public. The one thing Edison is known to have personally invented was the stock ticker.
