I debated writing this column about Microsoft's latest desktop initiative, but eventually my frustration and sense of impending doom won out. I think Microsoft is barking up the wrong tree with Plug and Play. This standard, which the software company announced earlier this month, has been in progress for about a year. Microsoft has been working in concert with a broad spectrum of PC industry vendors to define a standard set of APIs and software routines to detect add-on cards. The goal is to bring some sense to the current anarchy that we live in, especially when it comes to ISA bus machines. You'll need to upgrade three things on your equipment to make them P&P-ready: the operating system, the machine's BIOS, and the peripherals and their associated device drivers. That's alot of upgrading. Imagine being able to insert a card into such a PC, turn it on, and have it configure itself with the correct interrupt, DMA channel, and memory address. Imagine a PC that was smart enough to fix any configuration conflict before it booted, rather than just sitting there dumbly lost in the interrupt jungle never-never land. In short, imagine a PC that is as simple to configure as a Macintosh. It seems like such a Mom and Apple Pie issue: in these days where your support staffs are streched thin, it would be great not to have to spend hours adding a new card to an old PC. Indeed, just the other day my machine hung because I forgot to add a single "Int 2" line in my NET.CFG file for my network card. Without it, my machine booted and IPX drivers loaded, but it wouldn't connect to my server over my network. A simple matter of changing the jumpers on the network card and forgetting to add the corresponding line in the configuration file. A simple matter that took about 35 minutes to track down and fix. So I, like you, desperately want to be able to plug and play, and would like to see Microsoft et al. succeed in their efforts to get this standard adopted by everyone as fast as possible. However, Microsoft is going about it all wrong. While creating something that is strategically correct and an architecturally elegant solution, they have made the following tactical errors: First, they have ignored the installed base of DOS users. There is nothing in their plans for P&P to add support for this standard to DOS. And that's a darn shame, because DOS machines is what we all have got lots of, and continue to buy. Indeed, at the press briefing they demonstrated P&P on a machine running an alpha version of Chicago, the next version of Windows that doesn't run on top of DOS. Chicago is right now in the hands of a couple hundred people outside the Redmond campus. Second, they have given up on adding support for P&P to Windows 3.1, concentrating on Chicago and versions of Windows further in our computing future. And by Windows 3.1, I mean all of its cousins as well: Workgroups and NT. Intel stated that they will ship a driver to enable P&P to happen for existing Windows customers, and that Microsoft was perfectly happy to let them take this position and bat cleanup. And that scares me. Not to belittle Intel's efforts, committment, and abilities, but I really don't want to have to get my Windows software from Intel. Third, they have made the fatal mistake of working with their friends, and ignoring building lasting alliances with their enemies. Sure, you got to start somewhere and it is easier to build consensus when you use people that are part of your social or professional circle. But you get a warped picture of the world, and it is harder to build wide-ranging consensus down the road. I'll give you some examples. Sharing the podium was 3Com and National Semiconductor, not IBM's Raleigh Token RIng division. And the usual crowd of Intel sympathizers such as Compaq, Dell and AST, rather than the Power PC crowd or Motorola. Operating system vendors? Besides the home team, companies like Novell, SunSoft, and SCO were no shows. The sole representative from IBM said little and was from the PC Company, not the OS/2 Austin development group. And so forth. You get the picture. Now, I guess I shouldn't criticize Microsoft when Novell doesn't show up at their party. But nevertheless there is a certain hubris that the Microsoft representatives exhibited at the meeting. They really aren't interested in getting Novell's opinions, and would rather work among their own clan. Finally, they are about to make the same mistake that Compaq and AST did when they formulated EISA, and IBM when it came out with Microchannel. The world doesn't want a more sophisticated bus for the general desktop appliance. Especially one that adds an extra $1,000 cost on top of what is now a commodity item. Maybe we want high-performance network adapters inside specialized servers. Maybe we want better displays for our power users that can redraw at a rate more approximating a blink rather than a yawn. But these are minor portions of the overall PC marketplace. If P&P succeeds, EISA will be a dead duck. There may be a million or so machines extant, but how many of them really have more than one EISA card in them? And most of the EISA gang are already tooling up for selling you Yet Another Bus: PCI, VESA, whatever. Just trust us: it will be better and do more for you. Just junk what you've got now. And Microchannel is stalled as well: so there are seven million machines. So what? Even IBM is once again selling ISA PCs, and in fact darn proud of it (witness the ads for Ambra, its latest botched attempt at selling systems). So do you think I am all wet, or do you agree that the installed base of 120 million plus DOS users really could use this technology today? Let me know. If you'd like me to forward your reply to Microsoft, just say so in your email. I'd want so much to see Plug And Play succeed, but it will take a different set of tactics and different players to do so. And a different mindset: after all, if I really want Plug and Play today, I'll buy a Mac.
Click here to return to the
previous
page
David Strom
Port Washington, NY 11050 USA
US TEL: 1 (516) 944-3407