Google summer of code/netbsd mini-rant
dr
Herein lies a mini-rant:
http://www.netbsd.org/contrib/projects.html#kismet

"Improve on the Kismet design and implementation in a Kismet replacement for BSD."

Heres a hint, people:  The code is modular for a reason.  If you want to add support
for another platform that has rfmon, <b>talk to me for half an hour</b> and I'll
tell you where all the API hooks are and how they work. 

<b>Never mind the fact</b> that with netbsd moving towards Radiotap the last I heard,
Kismet will <b>already</b> work fine on it.

And heres the wonder of Open Source:  You don't like how something works in the 
code?  <b>Make it work better</b>.  Sam and Pedro have already contributed patches for
Kismet to work on FreeBSD and OpenBSD, respectively, and as soon as NetBSD uses
radiotap, it'll work there.  Total number of lines of code different to add OpenBSD
radiotap support?  About 550 lines.

Really don't like how things work in the code right now?  Again, <b>talk to me for
half an hour</b> and check out the newcore development branch, which contains the 
total rewrite to a completely modular, pluggable framework.

Maybe I just sound cranky about this, but it seems like a terrible waste of effort
to recreate 40,000 lines of code when you could make 500 lines of modifications
and maintain compatability with existing tools.  Why not take advantage of the 
corpus of existing code offered by open source.  Isn't that the point of
releasing the code in the first place?