Thanks for the reply, Chris. I understand and sympathize with the challenges of maintaining open source code or any other kind of volunteer work for that matter. I have some experience in X.25 (no "A") from long, long ago, and met Phil and Howie in the 80's at Dayton and I still have a TNC-2 somewhere in the basement.
This is an interesting discussion, and while not experienced with Go, nor having been a professional programmer since the 80's (I've learned C, SQL and Python since then, though), I'd like to see where this could be taken. The Go code is easy enough to read. The AX.25 ecosystem is a little confusing and I'd like to understand that before delving into code. It seems to me that the KISS solves a problem that no longer exists although I guess some people still like to use hardware TNCs. I don't understand how AGW fits in just yet. I found an AGWPE page and it seems rather full of itself with warnings about reverse-engineering and such, and I wonder how Direwolf, Soundmodem and Winlink Express got around that and what the limitations are. Seems to me that there is a bad smell coming from that direction, but maybe I'm not looking in the right places. Direwolf, on the other hand, appears to be very good, and if Wikipedia is to be believed, the only implementation of AX.25 V2.2, which should be superior for bulk traffic a la Winlink, although AX.25, like X.25, is well past its best-before date.
John, thanks for the pointer to LinBPQ. I'll give it a try. While I associate Telnet with the BBS days, I can see how it is useful for compatibility.
P.S. I noticed the FreeDATA modem project a week or so ago in my studies of OFDM, as I wanted to see how OFDM was implemented in FreeDV and what was going on with ARDOP. I think I have about 3 years of catching up to do on communications theory. I played with FreeDV a bit a few years back and may try again once my HF beam is fully operational again. My conclusions so far is that wide band communications needs to be embraced and that the FCC rules about symbol rates is a barrier to effectie HF data communications south of the border. The Canadian regulations simply give -26dB (of maximum) bandwidth limits on each band. 1kHz on 30m, 2.8kHz on 60m, 6kHz on other HF bands 160 through 12, 20kHz on 10m, 30kHz on 6 and 2, 100kHz on 125cm, and 12MHz on 70cm, and so on, with no restrictions on symbol rate . It seems like a lost opportunity for SDR transceivers especially with the processing power available these days without going to FPGA or DSP chips. But that's for a different discussion. A 48kHz sound card in I/Q mode can give about 48kHz bandwidth which the military uses on HF get to very respectable data rates. I don't see why hams can't do the same within licence limits.
In any case, I wish you luck with FreeDATA. And I wish the ARDOP folks luck with their OFDM work. I'd like to see hams get behind an open source modem that competes effectively with VARA, VARA FM, and Pactor. With a higher data rate we might get to the point where we could even accept the overhead of TCP and make full use of the bands and bandwidths available to us.