>	I believe that IRD ports, on the hardware level at least, show up as
>com ports on NetBSD.  As far as I know, there is no software on NetBSD,
>though there may be some on Linux which could be ported to NetBSD, which
>speak the IRDA protocol over the Infrared port.  The question is, do we
>want IRDA or BlueTooth as the first over-infrared protocol in NetBSD.

Errr, BlueTooth is RF, not IR AFAIK.

>Either would be nice, but my understanding is that both are a piece of
>work.  However, Two NetBSD boxes using the raw com port should be able to
>talk to each other via PPP without trouble via an Infrared port today with
>no software changes.

Actually, that's _not_ true (I speak from personal experience).

IrDA hardware has a few limitations; one main one is that it's all half-
duplex (typically the receiver is squelched when you're transmitting).
Another is that you have a non-zero turn around time between transmitting
and receiving (you negotiate it during the initial connection).

I _have_ logged in over IrDA hardware connections by running a getty on
the IrDA com port.  Once it was to chopps laptop at IETF, the other
time it was to a modified Shark.  I can report that running ppp over it
does not work (nor did I expect it to); it looks like the half-duplex
part was the killer.

When I had more free time I almost started writing an IrDA implementation.
It is actually very well documented (http://www.irda.org).  It will require
some work, but if you code it up as a finite state machine it should be
a simple matter of reading the FSM out of the documentation you can get
off of the web site.  Note: I never wrote any code; if you want to do
one, don't let my vaporware get in your way!

I did try to get specifications on BlueTooth, but it involved signing a
really horrible NDA that my boss wasn't willing to do (it also wasn't
clear that they would give us specs since we weren't hardware
manufacturers).

--Ken