On 11/14/2009 11:36 AM, Curtis Olson wrote:

> I suppose we could use some heuristic such as:
> 
> 4 character airport code that do not start with "K" or "P" use a leading
> zero, and all other airports omit the leading zero?  We could setup the code
> logic to be extensible if we find other countries that also tend to omit the
> leading zero.  That doesn't give us individual control over individual
> runways, but it might make things generally better than they are now?


That's the right question.  Let's see if we can quantify
the answer.

If we always omit the leading zero, we will be wrong
      37% of the time, approximately.  This is the
      current behavior.

If we always use the leading zero, we will be wrong
      the other 63% of the time, approximately.

If we use the K... P... ... heuristic, that improves things
      about 29%, so we will be wrong "only" about 35%
      of the time, approximately.  That is to say, there
      are quite a few US airports that are wrongly missed
      by the K... P... ... heuristic, and a few non-US 
      airports that are wrongly caught.  So this would be 
      only a rather marginal improvement.

============

If anybody is interested, I can provide a file "apt-state.dat"
that non-heuristically specifies which airports are in the
US -- and even specifies which state.  This would lower the
error rate rather dramatically.

By way of further improvement, observe that Air Force bases
within the US generally use leading zeros.  Google Luke AFB
or Dover AFB or Andrews AFB if you want to see some examples.
We can check the airport name (in apt.dat) and see if it
says "AFB".

Alas, it appears that some Navy facilities use leading zeros
(Miramar) while others don't (Lakehurst).  Eliminating this
source of error will take an hour or two of hand-work.

Finally, there are a handful of airfields in the Pacific that 
drop the leading zeros.  We can pick these up with the P... 
heuristic.

There may be some other check(s) that might be worth making,
but none that I can think of right now.

Bottom line:  It should be easy and straightforward to 
reduce the rate of leading-zero errors to negligible 
levels, i.e. to some rate very small compared to the 
rate of plain old errors in the database.


This is definitely worth doing.  Having the wrong markings
on a runway looks mighty peculiar, even if it is something
with no real significance, such as a leading zero.


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel

Reply via email to