I have some (usability) questions and comments regarding Netjack:
Why is it called jacknet_client instead of netjack_client (and it has
a zillion other names in the Readme, the source code and the web page
I would rename the -p parameter to -h (for host). I associate -p with
period or port not with peer.
Socket and port number have not exactly the same meaning. "-s <socket>
select another socket than the default (3000)" sound a little bit
strange. I would drop the parameter, you could add the port number to
the host name (-h hostname:3001)
-C could be renamed to -o (--outchannels) and -P to -i (--inchannels).
Is more constistent with jackd's parameter usage.
"-l <latency in periods> number of packets on the wire to approach" is
not clear to me. Is period the same as packet? Is it the period of the
local jackd or the remote jackd. Or do you mean frames?
What is "-f <downsample ratio>"? Has it always to be downsampling or
could it be upsampling? Should it be called resampling? And how do I
know which ratio I have to apply?
As reference the command line output from jacknet_client:
-n <jack name> - reports a different name to jack
-s <socket> select another socket than the default (3000).
-p <host peer> the hostname of the "other" machine running the jack-slave.
-P <num channels> number of playback channels.
-C <num channels> number of capture channels.
-l <latency in periods> number of packets on the wire to approach
-r <reply port> When using a firewall use this port for incoming packets
-f <downsample ratio> downsample data in the wire by this factor.
-b <bitdepth> Set transport to use 16bit or 8bit