On Mon, Nov 28, 2011 at 10:34 AM, Jim Reisert AD1C
<jjreisert@alum.mit.edu>wrote:
> In my opinion, having hardware/software to do this for you takes the
> "sport"
> out of Radiosport.
>
A long time ago, mid-70s to be exact, I was part of a research project to
see if a computer could be programmed to understand hand-sent Morse. From
a computer science viewpoint, Morse was an interesting problem because it
was much simpler than understanding speech, but could be used as a stepping
stone along the way to speech capability.
We picked a target environment of the typical 80M CW traffic net of the day
- lots of people checking in and out, a net control, stations moving up or
down to pass traffic, etc. Since I participated in the actual traffic nets
(EPA mostly), the idea was to program the computer to act the way I did,
which triggered all sorts of questions - like how do you manage to often
identify a specific station/operator just after a few characters, even
without hearing any callsign.
The computer system was built, and proved able to track the activity in a
net. It was especially cool to tell the program "Go find K3FIV", and watch
it as it methodically looked at each signal, and reported which one was
K3FIV, without ever hearing a callsign at al - just like a good opl. It
could "listen" to net activity, understand that two stations were being
sent off to pass traffic, follow and find them, listen to their exchange,
and then come back to the net frequency.
The next step, as I was leaving to work elsewhere, was to add the code for
the program to actually participate in a net - check in, pass traffic, etc.
But we were running up against resource limits. Even with a million-dollar
plus computer system that occupied much of a building floor, it couldn't
quite keep up with the real time nets at 30+WPM. The programming problem
was solvable, but the hardware wasn't up to the task.
Of course, that million-dollar computer of the 70s was a lot, really a lot,
less powerful than that old DOS box you threw out a few years ago. Today,
I can talk to the $300 tablet sitting in front of me and ask it to do
things - it understands my voice. Hardware has changed a lot in 40 years.
So, I think it would be straightforward now to write that old
Morse-understanding computer program again, but targeted toward contesting
rather than traffic nets. Contesting exchanges are much simpler than
traffic handling. With the right program, we humans wouldn't have to be
involved at all. It would be fun to write it, and to watch it play in
contests.
But, I think it would be wrong to use it. I like playing Radio.....myself.
73,
/Jack de K3FIV
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|