vtmatch passes all its tests now.  The next release of
vlfrx-tools isn't ready yet but the latest vtmatch is at

 http://abelian.org/vlf/tmp/vtmatch.c

which you can just drop into the build directory and do a
make install.

I played around with vtmatch on last night's VO1NA
messages and noticed some significant correlation
spikes at 08:30 09:00 and 09:30 but they failed to
decode.

Joe wrote:

> I'll try a longer message 39 characters, shorter symbols
> 0.5s and 8K19A starting 2000 UT.

Longer messages enable us to use longer lists in the decoder
without risk of false decodes.  We get a bit more coding gain.

VO1NA 2015-12-03/04 8K19A 0.5S 39C

 start  Eb/N0  BER    rank

 20:00   11.6   3.8%  0
 20:30   10.6   5.8%  0

--
Paul Nicholson
--