A suggestion:  the nntplink error reporting can make for a fairly
large log when network failures occur.  For instance, I now have quite
a lot of lines in my log reporting that zaphod refused to talk to us
:-) Any way to report the first instance, and then report them in
batches of 10 or so, or making the backoff faster?  The most
sophisticated solution is to report the beginning of the error
condition and the end, but that means that a successful operation has
to check if the last operation was an error. (moraes@cs.toronto.edu)

profile the code and add comments.

improve the error messages in link

when you get around to looking at your man page, it might be nice to the
default values for most of the options which have parameters. Like -C, -a, ie,
etc.

pull out the patches to news.daily so that the logfile aging won't screw up
everyone!

Add a new bit of log output saying how many articles nntplink tried to send
that didn't exist.  (jmalcom@sura.net)

Create a way to increase the debugging level (from none even) of a running
nntplink or a way to get it to dump its current view of the world (like
what you get when nntplink -d starts) or both?  Or create a way to force an
update to the .link file (when using batchfile mode).  (tale@uunet.uu.net)

add comments about how to interact with INN
