[kismac] Re: KisMAC 0.05e

At 12:34 AM +0200 2003/09/16, Michael Rossberg wrote:

>  kismac 0.05e2 will show the BSSIDs.

        Cool.  Glad to here it.  Will we be able to turn off the network 
name(s), and have just the BSSIDs?

>    i am still working on a good implementation of the algorithm.

        For help with good color selection, check out tools like Color 
Consultant Pro.

>  i will need to learn a little more on statistics for this (not too hard
>  i know, but i did not have this at school)...

        Median is easier than mean (average) and standard deviation.  For 
median, you just keep a sorted list of the measured values, and you 
take the middle number (for odd numbers of samples).  So, if you had 
99 measurements, the median would be the value stored at element #50, 
because 99+1/2=50.0.  For even numbers of samples, you average the 
two middle samples (100+1/2=50.5, so average the values at #50 and 
#51).

        This is actually a special case of percentile -- median is the 
50th percentile, saying that 50% of all measurements are equal to or 
above this number, and 50% of all measurements are equal to or below 
this number.  Other percentiles can also be interesting, but I don't 
drag you down that path.


        For calculating mean + sample standard deviation, there are 
existing formulas that exist on the net, but the math behind them is 
a bit more difficult to understand.  I think I've got some examples 
in perl I can share with you, if you like.

>  i did not program the traffic view, but as far as i can tell this seems
>  to be some kind of a bug. still needs a little more investigation...

        Ahh, okay.  I'll be glad to provide any help I can.

>  the problem that i see is, that this measurements will not be really
>  comparable, because the numbers that kismac shows, are depended on the
>  card and the firmware...

        I'm most concerned about measurements of the house with the same 
card.  Comparison between cards will be interesting, but the really 
useful information will come from double-ended network-based tools 
like pathload and pathrate.  I already have a test server in the 
basement on 100Base-TX, so that's one end point defined, and the 
other end point would be my wireless testing station.

        Signal strength (as reported by the drivers) will be one 
interesting measurement, but I'll also be looking at reachability 
(some cards simply may not have the power and sensitivity to cover 
the whole house), and network peak and sustained throughput.

-- 
Brad Knowles, <brad.knowles@xxxxxxxxx>

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
     -Benjamin Franklin, Historical Review of Pennsylvania.

GCS/IT d+(-) s:+(++)>: a C++(+++)$ UMBSHI++++$ P+>++ L+ !E-(---) W+++(--) N+
!w--- O- M++ V PS++(+++) PE- Y+(++) PGP>+++ t+(+++) 5++(+++) X++(+++) R+(+++)
tv+(+++) b+(++++) DI+(++++) D+(++) G+(++++) e++>++++ h--- r---(+++)* z(+++)

Other related posts: