The Emperor's New Clothes
Regarding Karen Forster's article "Corporate Insecurity, the Magic Mirror, and the Emperor's New Clothes" (March 2006, InstantDoc ID 49334), it would appear that what we are hearing from Microsoft is the same line we get from the government: "Yes, we have issues, but we are working on them and already have processes in place to correct them."
Unfortunately, Microsoft, like the government, can't be taken on its word. Only its actions count. Although I'm generally pleased with the changes Microsoft has made so far, I believe that once the company feels it has again locked the market down, it will go back to the old ways.
Microsoft has provided unbelievable compatibility from one OS to the next. For example, I had an application that was compiled for Windows 1.0, and that same executable ran on Windows NT 4.0. It of course took two companies— Microsoft and Intel—to create the full compatibility.
Because the UNIX world allowed universities and other developers to create ad hoc components for the OS, it has been impossible to develop software that will run on all flavors of UNIX. In fact, at one company I know of, the cost of migrating a C/Fortran program from one UNIX OS to another was estimated at $1,000,000 and would have taken from nine months to a year to complete. It was a huge application. However, the UNIX gurus swore up and down that they had used standard UNIX languages and compilers to create the application. The answer was as simple then as it is today—UNIX companies included features that locked their customers into UNIX products and hindered portability.
The Emperor still has no clothes.
Case in point: certification. Numbers are dropping. Companies don't want to pay for their employees' certifications because they're afraid that newly certified people will leave for other jobs. Individuals don't see pay increases that justify paying on their own for certification. Yet Microsoft is rolling out "new" certifications, thinking that if it renames a certification, it can fool people into thinking the cert is new and improved and that people will pay more for the same thing with a different name. And what genius is responsible for the idea that people who pay to take certification tests don't want to know where they're weak but only want a pass-fail evaluation?
How about Microsoft's brainstorm to make everyone go to 64-bit versions of its software? The company's reasoning here seems to be: Everyone wants new hardware, so let's make sure they have to get it. Then, let's leave the SQL management tools out of the 64-bit version and make customers purchase and run the 32-bit version if they want to manage 64-bit database servers.
The lunatics still run the asylum. And they want you and me to pay more for 64 bits of flawed new software so they don't have to fix the flawed old software.
KVM over IP
I thoroughly enjoyed Jason Bovberg's informative article "Buyer's Guide: KVM over IP Switches" in the February issue (InstantDoc ID 48825). When I read Dave Lynch's letter to [email protected] windowsitpro.com in the March issue stating that he can "do everything he needs to do and more" without resorting to KVM over IP, I thought, "Well, that's good for Dave, but apparently he's never had to remotely control a server during a power-on self test (POST)," for example by bringing up a Windows server in safe mode or controlling a server while it boots from a CD. You cannot do that remotely with any technology (e.g., RDP/Terminal Services, Virtual Network Computing—VNC, Secure Shell—SSH) that requires that an OS be currently running. Factor in the convenience of not having to dive into a client's data center at 3:00 a.m. to control a hosed server during POST, then maybe the ROI on the relatively high cost of KVM over IP switches can be favorably quantified.
Jason does allude to that benefit in the middle of his article ("It's hard to deny the convenience of being able to wield BIOS-level control over any number of servers or devices, all through your browser"), but I think this point was lost among the other details of KVM over IP switches that Jason presented. Regardless, thank you, Jason!
Reader to Reader to the Rescue
I want to thank Bret Bennett for his Reader to Reader submission "XP Power Scheme Fix" (November 2005, InstantDoc ID 47716). I had been having the same problem Bret described for several days. Thanks to him, I had it fixed in a matter of minutes.