I got e-mails by the dozen on the subject of my newsletter of two weeks ago, and I apologize for the delay in my replies: I spent the week of July 1 on the island of Catalina, off the coast of Los Angeles, with a troop of Boy Scouts in what I can only call a real adventure. No business trip has ever included a 5:30 a.m. wakeup call, after sleeping on an ocean beach, followed by launching an eight-man canoe into chest-deep surf at 6:30.
The most common question concerning my July 1 comments on IEEE 1394 (“FireWire”) plug-and-play performance was, “How could you not look at Windows XP?” I admit, that was an omission, and one that I acknowledged myself at the time: My decision to defer day-to-day experience with Windows XP is driven, though, by the nearly unanimous comments Ive gotten from eWEEK readers who are likewise holding off on letting XP into their environments. Their most-often cited reasons include substantial investments in code that still requires Win9x, and concerns about intrusive auto-update and other mechanisms that complicate configuration management.
I am assured by many of you, though, that XP is fully the equal of Mac OS X in its immediate recognition of FireWire devices—as I expected it to be, given the emphasis that Microsoft placed on this interface in its hardware guidelines for XP-targeted PCs. My point should not have been that Apple got it right, while Microsoft got it wrong, but rather that “FireWire-compatible” merely makes a statement about connectors and bits—not a promise of usability. Unfortunately, I cast my comments in terms that fueled the continuing Mac-versus-Windows feud, an unproductive pursuit that I wish I had resisted.
Many of you asked what I meant by calling Apples Titanium PowerBook “the least evolutionary” of the machines that I compared. My meaning would have been clearer if I had called it “the least legacy-encumbered.” The phrase that I used was meant as a not-quite-compliment on the ease with which Apple leaves behind such conventions as SCSI, serial and parallel ports, which we continue to see—and, in most cases, demand—on our Wintel machines.
On XP, Evolution, and Why DOS Must Not Die – Page 2
I found a fresh reason to think about legacy ports, and the value of direct control of hardware, when my son and I returned from Camp Emerald Bay to face a commitment to build a signaling system for a game of “Scout Trivia Jeopardy”. I thought wed merely be rigging up buttons and buzzers, but my wife and son persuaded me that wed spend the whole night listening to arguments about which team had actually buzzed first. We needed a system that would only recognize the first team to buzz in, while also blocking out any team that pushed its button before the question was fully read—with appropriate time-outs and displays.
A brief attempt to diagram a system of switches and relays convinced us that this would be hard. Credit my wife with being the first to say, “There must be some way to do this with a PC.” Before bedtime Sunday, we had prototyped a parallel-port interface that could handle five separate pushbuttons; by Monday night, wed built the thing (with my son getting his first experience in soldering miniature connections) and written most of the software in good old BASIC according to my sons first actual flow chart. And by Tuesday nights meeting, it was even playing the Jeopardy theme song on startup, and displaying the name (or names—it also handled ties) of the patrols that had earned the chance to answer.
As we were leaving the local Signal Electronics store with some of the hardware that wed bought to do this project, the clerk (who had overheard our discussion) warned us, “Dont try to do this with Windows 2000 or XP. They wont let you get at the hardware.” I found myself wondering if I need to treat our old DOS PCs as irreplaceable devices: Five years from now, will it even be possible to buy consumer hardware that enables this sort of hacking without drastic surgery? IBM, I see, still sells PC DOS 2000, the current version of the IBM PC DOS that is in my opinion the best retail version of DOS thats ever been offered—but Computer Discount Warehouse wont let me search its desktop PC offerings for any preloaded operating system older than Windows 95, although I can always load the latter in its DOS mode.
Some of your comments on my July 1 letter contrasted the promise of “cutting-edge technology” coverage with my reluctance to let go of what works today. Its a dilemma. The technologies that offer the greatest potential, though, seem to me to be in the domains of software and networks, not in the continual disposal and replacement of network nodes with more complex devices—and Id like to feel that each wave of technology gives me a superset of what I could do before, instead of constantly finding things that I cant do any more.