Because I heard some alarming things about uucp-throughput, I decided to
build a test situation. The result of the first test was quite surprising:

uucico waubel - (1995-11-29 22:32:15.13 346) Login successful
uucico waubel - (1995-11-29 22:32:15.88 346) Handshake successful (protocol 'i' sending packet/window 1024/16 receiving 1024/16)
uucico waubel leo (1995-11-29 22:32:16.07 346) Sending /usr/src/sys/arch/atari.org/compile/BOOT/netbsd.gz (333477 bytes)
uucico waubel root (1995-11-29 22:32:19.22 346) Receiving /var/spool/uucppublic/news.gz
uucico waubel - (1995-11-29 22:40:50.46 346) Protocol 'i' packets: sent 333, resent 1, received 432
uucico waubel - (1995-11-29 22:40:50.53 346) Errors: header 111, checksum 99, order 5, remote rejects 1
uucico waubel - (1995-11-29 22:40:50.74 346) Call complete (516 seconds 742632 bytes 1439 bps)

Both files were compressed using 'gzip -9', to rule out the compression
build in to my modem. It's a 9K6 modem, that was communicating with my atari
at 19k2. Although the number of errors was quite high, throughput was not
bad. I used a ZLRB_RING_SIZE of 4096 (standard 1.1 BOOT kernel) and
hw-handshake (hardflow true).

Comments?

Leo.