Python for air traffic control?
cribeiro at mail.inet.com.br
Tue Jul 3 05:50:04 CEST 2001
William, please don't take me wrong. I would like to make some comments
about the air traffic discussion. Let's not make this in a full blown flame
At 21:33 02/07/01 -0400, William Park wrote:
>Judging by his thin skin, I don't think it was a serious post.
I would not say this; I actually think it was the opposite. Air traffic
control is a *very* hard problem, and Russ's question was not related to
the actual algorithms for traffic control. Insted I think that he was
asking about the feasibility of using Python to develop a system that is:
- Real time. You can't stand to lose a single event.
- Non stop: 24 hours per day, 7 days a week, 365 1/4 days per year.
- Mission critical: people's live depend on it.
Considering this, your reply must have sounded as a troll to him. I'm not
an specialist on this, but I can tell you that the problem in air traffic
control is *not* how to map airplane positions using radar input; is how to
calculate paths and coordinate communications with dozens, or even
hundreds, of airplanes, while following *very* strict safety rules that
tell how to solve conflicts that arise all the time. You have to manage the
queues for takeoff and landing, keeping everything in sync to make the most
of precious time.
As for the flight path planning, please note that airplanes can't travel in
any place in space. You can't see it but there are several "lanes" in air
space, defined by the authorities. The use of the radio frequencies must
also be managed by the tower to make sure that there are not any undesired
interference between different airplanes. The system must also balance a
high degree of automation and the need for human decision. This is not the
place for dumb decisions.
In short, it is *hard* stuff, and it pushes our ability to develop high
quality systems. That's the kind of stuff that design-by-contract is
supposed to be used for.
More information about the Python-list