Hi all,
Watching the latest discussions on the list lots of thing accrued in my
mind so let me voice some of the stuff I think would be worth picking
up:
The fact that we have the internet is a result of standards that were
developed along the way. And that those standards were agreed upon by
the majorrity. That's why the internet is so good today.
Analogously, the Packet Radio with its ax25 specification is with no
doubt a standard averyone has to respect and lean to in order to
function in the 44.xxx.xxx.xxx subnet of tcp/ip and in the ax25 world.
We shouldn't waste or underuse what we've got so far.
In my humble opinion what we need is standards within the PR protocols
and Linux. Who should do that:
- authors of fbb, NOSes, tsthost, tpk, dpbox, baycom, etc.
- authors of the ax25 implementation in the kernel, MAILGW
- authors of flexnet, xnet, etc.
- sysops of gateways on a sporadic bases (the encap lines prove there
aren't that many of them anyway).
So defined forum would further set standards for forwarding protocols,
compressions, standards for programs like bbs (in the style of [ABFHM$])
and axioms of mail conversion from ax25 to smtp and vice versa.
After that each author(s) would adjust his/her sofware and develop
corrections to comply with the standards. I know I am idealistic here
and perhaps need to take my medication now but it seems to be a very
viable way to go farward. There are so many incompatibilities nowadays
with no clear clue how to circumvent them.
Here are my suggestions. Let me start big right at the outset:
- why not enforce an existance of a committee that would add a
compression specification to SMTP and POP3 standards? Alas the very
usage of them in the wireless hardware layer is a sufficient argument to
implement it.
An extention like that would benefit all of the MTAs and POPes.
Now in a little bit more of a mundane tone:-)
- I am not backing up the LZW. It really is a patented method. If we've
got the zlib library we should use it. Then auhors could insert it into
NOSes, MTAs and other future (or futuristic:-)) programs. The fact that
NOSes use LZW doesn't postulate adjustment of all to it. Polishing up
everything would require an amendment of one and the other.
- I wouldn't suggest that fbb should speak smtp or pop3. Unless it would
benefit one and only purpose: sending and receiving in ax25 only!!!!!,
with zero contact with tcp/ip. In this way one would send an ax25
messages via smtp, and receive bulletins and messages via pop protocol
(getheaders and retr #numberofbulletin).
- The above postulates some kind of standard compression although I see
no clear reason why should the pop and smtp be included in the IP. If
someone uses only TST that an address is not necessary and an exchange
could take place with different SSIDs and over a raw AX25 (i.e sr8ycb as
a bbs, sr8ycb-2 as a smtpax25, sr8ycb-3 as a popax25), and tcp/ip could
be used on the outside while forwarding between BBSes.
- Apart from that a Linux BBS should be sitting in the kernel and be
ready to listen & respond on different ports with the same IP (i.e 2525
as an smtp, 2323 as a telnet, 11010 as a pop, etc.), there's no sense in
wasting IP addresses - this is what we want to stay away from.
- A user who logs into a linux shell always gains an unrestricted access
to the internet if there are no firewalls. Lots of gateways do not have
firewalls but where there are we should take it into account. We could
voice that to the authors of kernel to implement a possibility to bar
access to a subnet for individual uids. A patch for kernel-2.0.36 would
also do well for those reluctant to hop on the newest 2.2 line.
- Gateways because there are what they are should have a guest account.
A security aspect of a guest shell account should here be discussed in
regards of what client programs should be available to guest (bbs, mail
to sysop, news). Moreover we should consider changing "login" or
something else so that guest access from outside of a 44 subnet be
impossible!!
- shouldn't we dust off "axspawn" and see in real life what it could do
in this regard? Moreover we should consider an automatic creation of
accounts (only from 44 subnet, with call sign autentication in the way
Tnos had it) for sysops who would like it made available.
- coordinate nos, inn and nntp.
- Correct kernel ax25 "reaction" where it is outrun by flexnet
- I think a Ham radio operator should enjoy sending mail form the
internet to an ax25 world with the capability of an automatic address
conversion after some kind of autentication and callsign verification in
a standard mail software.
The same could be true of bulletins, i.e as a one way discussion list
(that, by the way, is a future feature of mailgw). How to do it? - we're
here to establish that.
After the whole work is done we should expeditiously start a campaign to
convert to the new standards, convince users to do the same:-))))))
After all there aren't that many gateways and bbses that it would take a
major revolution to adjust!! We could do it once and swiftly but this
time with the peace of mind years to come.
That'll be it from me.
What do you guys think?
73 de sp8qed
P.S. I would like to send thanks to Benedict KB2QZV for helping with
good translation of this letter.
