I was in a meeting today where I actually
heard this. “We have chosen this path
because software is easier to fix than hardware.” I had to suppress a chuckle, it was a serious
statement made by a serious engineer who was trying to explain why his radar
system didn’t work, and how they were going to fix it.
I was drawn back, against my will, to
1984. I had just returned to Special
Operations after four years teaching new officers to be Navigators in the
AF. The base was abuzz with the rumor
that a new airplane, the MC-130H, would be delivered by the end of the year and we were on a
track to completely modernize the Special Operations MC-130E inventory. The government was buying new C-130H aircraft
from Lockheed, IBM was putting on a new fully integrated glass cockpit, and
Emerson Electronics was building a brand new terrain following radar that was
going to be 100% better than that old Texas Instruments system on the MC-130E.
I left Hurlburt Field in mid-1986 and the
promise of the MC-130H was just around the corner, it would deliver in late
1986, just as soon as they fixed one or two little radar problems. For the next four years the MC‑130H was to
deliver in six months, just as soon as they fixed that one little radar
problem. It was always a newly
discovered, easy to fix in software, radar problem. Meanwhile the 22 C-130’s the Air Force had
bought piled up in an airfield in central Texas, waiting to be converted. They sat there so long that when they started
up the engines two of them almost blew their wings off.
So here I am 25 years later listening to
that same radar company tell me how simple it will be to fix the problems they
have caused with the changes to their radar system. I guess there really is symmetry to
life. A company can change names, but
the same screwed up excuses seem to survive.
The truth is software appears to be cheap
and simple, but can get really complex and expensive very fast. Once you commit to writing software it is
like looking at an iceberg. You see 10%
of what it will really involve.
Hardware on the other hand seems expensive, but it is really simple to
understand how to fix something that isn’t working. For example, take your home computer. If the hard drive crashes it is pretty easy
to figure that out and it may be a bit expensive to replace but a new one will
slip right in. If you have a software
problem that keeps causing your computer to crash you can spend hours and hours
trying to figure out why. It can be as
simple as how the applications work, but you won’t know that unless you can
read the code.
1 comment:
I know the hands on code... righty tighty, lefty loosie...
Post a Comment