LETTERS TO THE EDITOR | GCN
I find I'm not satisfied with your conclusion that Microsoft Windows NT bore no fault in the failure aboard the Yorktown ("Navy: Calibration flaw crashed Yorktown LAN" [GCN, Nov. 9, Page 6]). While these days the statement that NT was not the cause of the failure may be true, using some strict definition of the word cause, NT certainly failed to provide protection from a simple application error.
I find Im not satisfied with your conclusion that Microsoft Windows NT bore no
fault in the failure aboard the Yorktown (Navy: Calibration flaw crashed Yorktown
LAN [GCN, Nov. 9, Page 6]).
While these days the statement that NT was not the cause of the failure may be true,
using some strict definition of the word cause, NT certainly failed to provide protection
from a simple application error.
If I recall correctly, back in 1970 when I was running Fortran programs on an IBM S/360
while attending college, an error in data could cause my programs to abort. But it
certainly did not crash the whole system and make it unavailable to other users. Since the
early 1970s operating systems such as OS/360, Multics, Unix and others have protected the
integrity of the system in the face of wayward applications.
Fuel valve software causing uncontrolled growth of a databasewhat brand database
was it, anyway?would not be sufficient reason for a LAN and 27 high-powered remote
units to crash.
Some feel that NT still seems, all too often, to present a blank blue screen. As an
example, I have a laptop running NT that provides a blue screen at all attempts to search
for a location using Microsofts own Streets 98 software.
Name withheld
Editors note: The writer works for the Air Force.
In response to your story, Navy: Calibration flaw crashed Yorktown LAN:
Without either one of us knowing all of the relevant data, it still boggles the mind that
any divide-by-zero error on Microsoft Windows NT would cause a system to crash, let alone
27 of them.
I feel that this article is not much more than the higher-ups being technically snowed
into believing that it was not any fault of the computer, and thus it must have been human
error. I dont care what operating system, computer or application Im using, I
should be able to type in a zero and expect the computer, let alone an entire LAN, not to
crash, especially if that zero is to represent a closed valve!
Im glad that changes were implemented, especially with a military system where
the lives of our nations guardians are at stake.
Gil Young
Corporate network engineer
Coleman Research Corp.
Orlando, Fla.
About your editorial, DOD retreats on Web [GCN, Oct. 12, Page 20]: I concur completely with your comments
regarding the government letting fear force us to roll the clock back.
Overreacting to the new Defense Department Web policy by arbitrarily shutting down Web
servers is clearly a step in the wrong direction. Well said.
Maj. Jim Cummiskey
Marine Corps Mobile Computing Project Office
Camp Pendleton, Calif.
In response to your editorial, Pay is not the only way [GCN, Oct. 26, Page 18]: Thanks for pointing out that to be
sure, pay levels are too lowvastly so, in many instances.
However, the gist of the editorial and its title seems to imply that other things could
help retain capable employees in lieu of salary. Most managers wont bother to read
the column before drawing that conclusion from the headline because it makes them feel
better. Such a conclusion would be erroneous. Until the pay situation is rectified,
quality of life wont matter.
Name withheld