Re: [Tlf-devel] Autosend feature - first tests

[Thomas Beierlein]

Hi Martin,

Am Tue, 12 Nov 2013 11:05:42 +0100
schrieb Martin Kratoska <address@hidden>:

> Hi Tom et al,
> 
> Dne 12.11.2013 10:07, Thomas Beierlein napsal(a):
> > Hi Martin,
> >
> > Am Tue, 12 Nov 2013 00:33:09 +0100
> > schrieb Martin Kratoska <address@hidden>:
> >> there is something strange. tlf does not see now hamlib, so no
> >> radio control :-(
> >>
> > Maybe you forgot the '--enable-hamlib' in the ./configure stage?
> 
> Of course. Dumb me. A senior moment, apologies.
> 
Never mind. I know such moments myself :-).

> >> The Autosend option works but not exactly as meant. The auto
> >> sending begins with the character count as set with :char command,
> >> this is OK. Setting :char to 4 actually moves the cursor to the
> >> exchange field while tlf is sending 2nd character.
> >
> > That depends a little on your typing speed. But you are right. It
> > moves to the exchange field as soon as you stop typing.
> >
> > The reason for this behavior is that we can not do the same
> > handling as in TRLog. As TRLog sends the CW characters itself
> > (integrated keyer) it knows exactly how far it is with sending the
> > already typed in characters and when it catches your typing.
> 
> OK, I know, it is very hard task if tlf does use an external device.
> My setup includes the K1EL's Winkey USB (the obsolete ver. 2.2) and
> the Winkeyer daemon by OK2CQR
> http://www.cqrlog.com/?q=node/149
> 
> (BTW what about integration of Winkey USB support into tlf code? This 
> excellent keyer is widely used among contesters/DXers using Windoze 
> software, it has also dual output/PTT facility making SO2R easier.
> The PTT output not tested here yet because my K3 with KPA500 has
> excellent QSK properties and absolutely trouble-free T/R switching).
> 
Hmm. As far as i know the interface (via unix socket) is the same as for
cwdaemon. What are the advantages of an integrated winkeyer daemon?
Don't get me wrong - I am not against the idea, but want to have some
idea about pro and con of that solution.


> > For us using cwdaemon, the usbdaemon or Andys winkeyer daemon means
> > we can send out the characters to the keying daemon but get no
> > feedback how far it is with sending and therefore we have no
> > information when the daemon catches your typing.
> 
> BTW what is Andys winkeyer daemon? Where can I get it?

He wrote shortly in the thread about it. Here is the link he gave:
http://g4kno.com/software_it/jWinkeyServer/jWinkeyServer.html

> ... 
> > Any ideas how to come around that are welcome.
> >
> Now, after some 3 hours of thorough testing at speeds varying fron 18
> to 45 WPM I am pretty sure that my other thoughts were false and 
> over-complicated. The main fact is that the typing speed is almost 
> constant, independent of the keying speed. Again, the timeout of 0.7s
> is just fine.
> 
> The Autosend option now seems solved, at least for me.

Ok. So we have a first working solution.
Fred mentioned a configurable timeout (I am not sure if it is needed,
but let us see). There is another idea I want to try out, but I do not
know when there will be time for it.

So I think I will merge the autosend feature in the master branch and
make it default.

> >> Therefore the op has not
> >> time till 4th character is finished, here I have time actually only
> >> till 1st character is completed.
> >
> > Not quite. As described above you have 0.7 seconds for each
> > character you want to type.
> >
> > 73, de Tom DL1JBE
> 
> BTW many thanks for another very good version, Tom! Excellent work!
> 
Yw.

73, de Tom DL1JBE.

> 73,
> Martin, OK1RR



-- 
"Do what is needful!"
Ursula LeGuin: Earthsea
--