[drats_users] Possible Who Is Online feature request

[Dan Smith]

> I agree.  Bad idea.  Wasn't thinking about lots of stations.

I try to anticipate such things since I'm the guy that the admins
would (unfairly, IMHO) hold accountable for such a thing :)

> I was watching that with interest, but I'd be more interested not in
> something that changes packet size, but that shows an
> "instantaneous" view of error rate.

Yeah, that's something the radio needs to provide, I think.  Unless I
add actual error correction to the packets (as opposed to just a
checksum) I can't really determine what the bit error rate is.  Adding
ECC to the packets would make it significantly slower when your
connection is good. 

> The problem (usually) is not that larger packets get into trouble (I
> don't think), usually the stations are fixed (around here anyway)
> and they simply aren't solid enough "data copy" into the repeater to
> reliably do D-RATS.  Usually if you're "in" the repeater, you're
> "in" and you can send short or long packets as you see fit, from
> what I'm seeing here.

Yes, that's the typical case.  Your repeater might be good enough that
you really do fall off that cliff everyone loves to show in their
powerpoint slides.  When you're on simplex with less impressive
antennas, a tree swaying in the wind (or something) can put you in a
much more middle-of the road situation, signal-wise.

> Users often don't "believe" those of us who are solid into the
> repeater that this is their problem ("But I can talk just
> fine!")... you know the drill.

Sigh... only too well :)

> What would be useful would be a way to send a known bit-pattern in
> small packets continuously (like loopback tests in telco?) to a
> strong station and have that station tell you over a period of a
> minute or so (wouldn't want to do this very often... very annoying!)
> when you're setting up a portable station or whatever, if your
> signal is "good enough" to even do D-RATS in the first place.

Yep, this is a good idea, and one I used a lot in the early days to
figure out the banned characters and characteristics for each radio.

Sending A-ZA-Z-AZ a few times would let you check how much of it is
getting killed.  The problem is, it's the kind of thing you'd have to
coordinate over voice with the other side, and you'd need a clear
channel (clear of incidental GPS traffic too) to really do it well,
given no reliable packetization mechanism.

It gets sticky quick, and if it's unreliable, it's hard to trust it.
So, I've put it off thus far :)

> Even cooler... would be to teach the Gateway to do it... but that'd
> probably require some fancy coding by Robin, or at least assistance
> from him in using his sniffer to drive an application that could
> reply on-channel.  Maybe I could figure out some way to hack
> something together using the sniffer and the voice
> announcements... hmm... doesn't sound particularly easy... "Error
> rate, 0 percent", etc.  That's probably "overkill" though.  Another
> participating D-RATS station could do it.

Yep, there's a lot I'd like to be able to do on the gateway, but I
need some cooperation from the current landlords there :)

You can, of course, use the broadcast-text function in D-RATS and set
your radio to hit the echo module.  You have to eye-parse and
eye-calculate the error rate, but sometimes the visual is more telling
than a number :)

> Another thing to think about for much later down the road... for
> simplex... a way to "designate" a path... Station A knows they can
> reach WY0X, and they know Station B can also... but that's starting
> to sound like Packet AX.25 routing.

Yep, another thing I've thought about adding.  I have some plans to
implement a part of it (the unintelligent part) actually.  However,
the automatic path discovery part is not only hard, but it's also
something that can seriously interfere with voice traffic, especially
when I can't tell that there's voice traffic happening from the other
side of the radio.

So many people want to use it on a voice repeater, I hesitate to give
it a mind of its own at times for fear of the blacklash :)

Good thoughts and suggestions regardless!  More discussion on this
level will only help :)

-- 
Dan Smith
dsmith#danplanet.com, s/#/@/
www.danplanet.com
KK7DS