In-Depth

December Response Time

• 12/01/1998

In the article "Y2K Compliance Testing and Verification" in the August 1998 issue of ESJ, the author writes, "In Warren, Mich., the first Year 2000 lawsuit was filed when a retailer claimed he suffered significant financial setbacks when his system could not accept credit cards expiring after the year '00.' " There is no mention of the date of this lawsuit; however, it probably wasn't the first. I was involved in a suit in January 1989.

Our data backup and migration product attempted to extend the expiration date on a number of tapes containing archives to a Year 2000 date. The tape system promptly scratched the tapes, allowing them to be written over. This did result in a lawsuit against the vendor of the data backup and migration product, which was settled out of court for a modest amount.

I wish I had been smarter at the time in recognizing the nature and extent of the problem. I might have had something to show for my experience, other than a dubious "first."

I found the article "The System Sort" (September ESJ, page 99) to be quite an interesting way to age data in existing files. Page 103 lists an example of changing the year portion of a date in a file from "96" to "99." This will work correctly for all dates in 1996, except for one: February 29. Using the example listed on this page would generate an invalid date for 1999, since 1999 is not a leap year.

The article still has a lot of merit and good use of the Sort program. Caution must still be exerted when modifying dates within files to ensure the validity of the dates generated.

Your September editorial made me think that IBM had purchased. Yes, the mainframe's still alive and still a force in enterprise computing, but it's a declining force. What other product line measures itself in MIPS shipped? How many UNIX MIPS or NT MIPS shipped last year? The truth is that mainframe revenues are now about half of what they were six or seven years ago, and unit shipments are similarly down. The mainframe is now well into the first five years of a permanent 15(-year) decline. And why? For pure, economic reasons. Mainframes today are built from the same technology (CMOS chips) as SUN, HP or Compaq servers. IBM charges about $6,000/MIP for these, the others charge about $600/MIP and deliver equal performance. Why would you pay 10 times as much for anything?

Regarding the John Cleese humor column from the June issue: Didn't I read this same column a few years ago? I can't recall if ESJ (I've been a fan ... since 1985) had a strangely similar piece.

As I recall, Mr. Cleese had spent some time and energy building a management consultation firm; I was enamored of the previous column, because it so well fit my business climate at the time, as well as showing Mr. Cleese's grasp of topics, so apparently orthogonal to his previous career in entertainment.

I do wish I could place my hands on the previous John Cleese column. Perhaps your library is more encompassing? In any event, the column is just as pertinent today as seven or eight years ago.

John Cleese has never appeared in ESJ, however, while Editor of our sister publication, HP Professional, I asked Mr. Cleese to contribute a guest column on managing unproductive personnel. (March 1997, page 48). Due to the overwhelming response to this column we asked Mr. Cleese to share his keen-witted observations with ESJ readers on a similar subject - preserving order amidst change in the workplace. We found the response to Mr. Cleese equally as enthusiastic.

Charlie Simpson

Thank you for your comments about NT and UNIX, in your "Both Sides Now" editorial [appearing in the December 1997 Platform Decisions supplement at www.platformdecisions.com]. I take a special interest in both Linux and NT as I study both platforms.

Discovering stories like yours are quite rare, and I would like information for a source on the www to obtain such data on the rivals: NT and UNIX. Good writing.