The Microsoft standard is anything but

COMMENTARY--If you work in an office in the United States in 2001, Microsoft products are almost certainly standard issue. That doesn't make them true standards, however, despite the fact that Microsoft does everything it can to set, rather than follow, standards. Arguably, Microsoft's control over standards sets it up to both dominate and, one day, to fall.

True standards are approved by standards organizations like the IEEE, W3C, and ISO. Microsoft generally embraces standards but also extends them with proprietary features. Thus, Microsoft products typically interoperate with standards-based programs from other vendors, but they interoperate better with other Microsoft products. This gives IT managers an incentive to standardize on Microsoft.

For example, in my company, we use Exchange e-mail servers. Exchange has its own protocol for message retrieval, similar to POP or IMAP. Fortunately, our administrators have enabled POP support on our Exchange servers, which lets us use the standards-based mail client of our choosing. However, those using POP aren't integrating with the shared calendar. Managers therefore pressure all users to get on the same platform, so we can all work together most effectively.

This is exactly the effect that Microsoft hopes for, and it happens at a lot of companies. Consider a pal of mine who wrote large parts of Emacs, a popular "text editor/kitchen sink" in Unix environments. He had been on the Internet since 1982, and was using tools like Emacs, procmail, and gnus to manage and categorize thousands of mail and news messages per day. Yet, when his coworkers sent him an invitation to a meeting, he couldn't click on the little button to "accept" and have the event automatically added to his public calendar. They made him switch to Outlook. (Perhaps not coincidentally, he left the company a short while later.)

For its part, Microsoft requires all of its dial-up MSN Internet services customers to use Outlook. There may be some feature-ish justification for this, such as improved SPAM filtering. But from some angles enforcing email client choice through an ISP doesn't smell too different than enforcing default Web browser choice through an operating system. It's the same kind of hi-jinx that ran Microsoft afoul of the Sherman Act and its interdictions against anti-competitive corporate behavior. After all, every other email provider on earth abides by the standards--why can't Microsoft?

For that matter, why can't Microsoft build its shared calendar around open standards such as iCal or "shared source" the Outlook calendar spec with other developers?

Microsoft is not the first mega-corporation to attempt to control standards through sheer size and influence. IBM was the first 800-pound gorilla here. It made the PC, which had been just a toy for Heathkit hobbyists, into a business tool, and owned the whole PC hardware market until clone makers came along. It then tried to take back the market by redefining the PC hardware standard around Micro Channel Architecture. It was too late. Consumers stuck with ISA and clones.

The Internet revolution brought another 800-pound gorilla, or should we say Mozilla. Believe it or not, relatively few people gave Microsoft much of a chance of catching up with Netscape's lead in Internet software. One of Netscape's most powerful weapons in the battle for standards control was extending HTML. Its proprietary tags gave Netscape products more functionality than you could get with purely standards-compliant offerings, and Netscape encouraged people to add little colophon tags that said, "This page best viewed on Netscape." There was a catch, though. Not all of those hurriedly implemented tags wound up making it into the standard. If you work as I do for a company that built in-house applications around the <layer> tag and other apocrypha--until late 1998 Netscape was the only cross-platform browser, after all--you're now stuck with Netscape 4.x until you can find time to rewrite all your pages.

From these examples, it's clear that standards enforcement can be a tricky business. I'm not saying that Microsoft "standards" will disappear any time soon. However, if I were an IT manager, or especially a corporate development manager, I'd look upon any standard that isn't really a standard as a potential wolf in sheep's clothing. I'd be chary of getting locked into any standard that isn't one--even if it's a Microsoft standard.

What do you think? Is Microsoft really in charge of computing standards, or do the standards organizations still hold the real power? Let me know in the TalkBack below.