UUCP on eternal september

[Paul Edwards]

Hi.

I'm not familiar with the UUCP protocol,
and the Eternal September website makes
no mention of it (as far as I can see)
except in the "news".

So - it may be working but I'm incorrectly
expecting something unreasonable.

Here is what I get:

kerravon@kerravon-pc:~/Downloads$ telnet uucp.eternal-september.org 540
Trying 95.217.65.141...
Connected to uucp.eternal-september.org.
Escape character is '^]'.
login: (i typed my userid that works fine for nntp)
Password: (i typed my password that works fine for nntp)


Can someone confirm whether UUCP is still
a thing on Eternal September, and if so,
is there something else I need to do?

If I failed to obey UUCP protocol, that's
fine. That's what I'm going to start looking
into. But if it doesn't exist, that won't help.

Thanks. Paul.

[VanguardLH]

I've never tried using ES for UUCP (Unix-to-Unix Copy) file transfers.
ES is about articles in Usenet via NNTP.

https://en.wikipedia.org/wiki/UUCP#File_requests
(c.1979)

https://en.wikipedia.org/wiki/Network_News_Transfer_Protocol
(c.1986)
Usenet was originally designed based on the UUCP network, with most
article transfers taking place over direct point-to-point telephone
links between news servers, which were powerful time-sharing systems.
Readers and posters logged into these computers reading the articles
directly from the local disk.

As local area networks and Internet participation proliferated, it
became desirable to allow newsreaders to be run on personal computers
connected to local networks. The resulting protocol was NNTP, which
resembled the Simple Mail Transfer Protocol (SMTP) but was tailored
for exchanging newsgroup articles.

You're using Thunderbird which has no UUCP support. Tbird didn't show
up until 2003 (as Minotaur a year before). That was long after the UUCP
era, and after NNTP was established.

Are you logging using telnet in a command shell, and trying to enter
UUCP commands via a telnet shell? UUCP is for file transfers, and also
for remote command execution (and why it needs root privs). Most times
when I've seen UUCP used, it was to run the 'uucp' command to copy a
file from one *NIX host to another.

I have an ES account. However:

telnet uucp.eternal-september.org 540
login: <myusername>
password: <mypassword>
Connection to host lost.

So, the UUCP connect fails, and the server disconnects, when using the
login credentials for my ES account. NNTP works just fine with ES using
the ES account login credentials.

telent news.eternal-september.org 119
200 <statusInfo>
authinfo user <myusername>
authinfo pass <mypassword>
281 Authentication succeeded
list
<wholeSlewOfNewsgroups>
quit

Doesn't look like ES does UUCP. I doubt it ever did. ES started in
2009 which evolved from Motzarella started in 2006. That's long past
when the UUCP relic fizzled with later protocols (SMTP, NNTP) replacing
it, and with analog dial-up replaced with always-on cable/DSL/sat/GSM
connects (although UUCP can be adapted to use the later communication
venues). UUCP does remain a resilient store-and-forward protocol, like
using it to upload readings to a server from remote weather stations.

I'm not the ES admin (Ray Banana). Wait for him to give a conclusive
response. I've not heard ES was ever a file storage service, and I
doubt Ray would ever allow remote command execution on his servers.

[Paul Edwards]

On 26/12/23 16:17, VanguardLH wrote:
> Paul Edwards <muta...@gmail.com> wrote:
>
> I've never tried using ES for UUCP (Unix-to-Unix Copy) file transfers.
> ES is about articles in Usenet via NNTP.

You can (in general, not commenting on ES),
use UUCP to get news articles.

That's the traditional/sophisticated way to
get it in bulk for offline viewing/responding.

I came up (I'm 56) via Fidonet, not Usenet,
so didn't have much to do with UUCP, other
than I knew it as the "rival".

> You're using Thunderbird which has no UUCP support.

I will be writing my own software, to replace
pdpnntp, which runs under PDOS/386 (http://pdos.org)

> Doesn't look like ES does UUCP. I doubt it ever did.

Well, this is from the front page of the ES website:


2011-05-26 02:59:03 Server names

Due to recent incidents, please note that you should only use
"news.eternal-september.org" as the server name in your client
configuration. IP addresses and other server names may change any time
without prior notice and hence become unavailable for usenet access.
Exceptions are "reader80.eternal-september.org" for access on port 80
(NNTP), "reader443.eternal-september.org" for access on port 443
(NNTP/SSL), and "uucp.eternal-september.org" for UUCP access on port 540.

> 2009 which evolved from Motzarella started in 2006. That's long past
> when the UUCP relic fizzled with later protocols (SMTP, NNTP) replacing
> it, and with analog dial-up replaced with always-on cable/DSL/sat/GSM

I'm after a solution that works for not-always-on
situations. E.g.

https://en.wikipedia.org/wiki/Sneakernet

Bhutan
The Rigsum Sherig Collection project[6] uses a sneakernet to distribute
offline educational resources, including Kiwix and Khan Academy on a
Stick,[7] to hundreds of schools and other educational institutional in
the Kingdom of Bhutan. Many of the schools in Bhutan have computers or
IT labs, but no Internet connection (or a very slow one).[8] The
sneakernet, facilitated by teachers, distributes about 25 GB of free,
open-source educational software to the schools, often using external
hard disks.


This (Bhutan) may or may not be the target
market for PDOS/386.

Regardless, I want to have a technical solution clear in
my mind, and that is get data from Usenet, probably in
bulk via UUCP, write some gateway software to convert it
into FTN format, preserving the Usenet headers
with x'01', copying across the X-Comment-To: (ditto
in reverse) to the "to" field, where tobruk can be
used, then msged as the reader, micro-emacs as the
editor, and then have the reverse process.


BTW, as to UUCP being outdated - I saw a message
in another ES group saying that newsgroups were
getting "revived". I wasn't aware of that. I've
been using them for decades, and I'm now expanding
the software (including pdpnntp) for my own interest,
but I wasn't aware there was a revival. Did something
happen?

I can remember reading something from Microsoft saying
that ANSI terminals were their new focus for portability
too. I'm doing that too. Even on the mainframe with
EBCDIC ANSI (see z/PDOS from same spot). Seems similar
to me - old tech revival?! I mean - sure - I'm doing it
too - but I'm really weird. I'm surprised anyone else
is interested. I've spent a lot of effort on 16-bit too.
I bought Microsoft C 6.0 (brand new) and that arrived
a couple of weeks ago.

BFN. Paul.

[yeti]

VanguardLH <V...@nguard.LH> writes:

> Are you logging using telnet in a command shell, and trying to enter
> UUCP commands via a telnet shell?

That's ok, while I'd prefer `netcat` to test UUCP-over-TCP540.

> UUCP is for file transfers, and also for remote command execution (and
> why it needs root privs). Most times when I've seen UUCP used, it was
> to run the 'uucp' command to copy a file from one *NIX host to
> another.

The `uucp` command is at a different level. Under the hood some UUCP
nodes can be defined as being reachable by different methods. Using
TCP540 is one of them. You would not notice the difference unless you
start debugging the connection or remote system definition. Somtimes
you might hear a modem chirp, sometimes it just would work silently.

[yeti]

VanguardLH <V...@nguard.LH> writes:

> Are you logging using telnet in a command shell, and trying to enter
> UUCP commands via a telnet shell?

That's ok and so is `netcat` to test UUCP-over-TCP540.

> UUCP is for file transfers, and also for remote command execution (and
> why it needs root privs). Most times when I've seen UUCP used, it was
> to run the 'uucp' command to copy a file from one *NIX host to
> another.

The `uucp` command is at a different level. Under the hood some UUCP
nodes can be defined as being reachable by different methods. Using
TCP540 is one of them. You would not notice the difference unless you

start debugging the connection or remote system definition. Sometimes

[Marco Moock]

Am 26.12.2023 um 15:13:57 Uhr schrieb Paul Edwards:

> I'm not familiar with the UUCP protocol,
> and the Eternal September website makes
> no mention of it (as far as I can see)
> except in the "news".

In the innreport files some UUCP peers are named:

http://feeder.eternal-september.org/index.php?showpage=stats

[VanguardLH]

Marco Moock <mm+s...@dorfdsl.de> wrote:

> Paul Edwards:
>
>> I'm not familiar with the UUCP protocol, and the Eternal September
>> website makes no mention of it (as far as I can see) except in the
>> "news".
>
> In the innreport files some UUCP peers are named:
>
> http://feeder.eternal-september.org/index.php?showpage=stats

I saw only 1: uucp.gnuu.de.

Not sure how to read the INN stats, but looks like only 5.7% of the
offered articles were accepted, but don't know if that was from ES to
gnuu or from gnuu to ES. In any case, very low traffic between ES and
gnuu. Plus the peering was over port 119 (NNTP), not 540 (UUCP).

[Marco Moock]

Am 27.12.2023 schrieb VanguardLH <V...@nguard.LH>:

> Marco Moock <mm+s...@dorfdsl.de> wrote:
>
> > Paul Edwards:
> >
> >> I'm not familiar with the UUCP protocol, and the Eternal September
> >> website makes no mention of it (as far as I can see) except in the
> >> "news".
> >
> > In the innreport files some UUCP peers are named:
> >
> > http://feeder.eternal-september.org/index.php?showpage=stats
>
> I saw only 1: uucp.gnuu.de.
>
> Not sure how to read the INN stats, but looks like only 5.7% of the
> offered articles were accepted, but don't know if that was from ES to
> gnuu or from gnuu to ES.

Usenet servers only accept an article if the following requirements are
met:

At least one of the group the message is posted to exists on the server.
The message is not rejected by filter mechanisms.
The article doesn't exist on the server (that should be the most
relevant reason for refused articles).

> In any case, very low traffic between ES and
> gnuu. Plus the peering was over port 119 (NNTP), not 540 (UUCP).

See here:
http://news.eternal-september.org/stats/news-notice.2023.12.28-04.00.01.html#batcher_elapsed

It seems to come in at the news. server and not the feeder.

[Paul Edwards]

On 28/12/23 16:25, Marco Moock wrote:

>> In any case, very low traffic between ES and
>> gnuu. Plus the peering was over port 119 (NNTP), not 540 (UUCP).
>
> See here:
> http://news.eternal-september.org/stats/news-notice.2023.12.28-04.00.01.html#batcher_elapsed
>
> It seems to come in at the news. server and not the feeder.
>

I tried:

telnet news.eternal-september.org 540

but got:

telnet: Unable to connect to remote host: Network is unreachable

BFN. Paul.

[Paul Edwards]

On 28/12/23 17:58, Paul Edwards wrote:

> I tried:
>
> telnet news.eternal-september.org 540

It occurs to me (I'm from Fidonet, not Usenet) -
is this considered to be "peering" and thus not
available to "normal" users?

In Fidonet end users could become "points" off
an ordinary node, and that is what I was expecting
UUCP to be the equivalent of.

But given that a "machine name" is part of the
process, maybe they never had what Fidonet had,
and maybe that is partly why Fidonet came up
with something more appropriate/practical in the
first place?

I must say - I wasn't expecting to come here and
find no (ordinary) UUCP users at all. No demand
for bulk, efficient message transfer?

BFN. Paul.

[Marco Moock]

m@localhost:~$ telnet uucp.eternal-september.org 540
Trying 2a01:4f9:4b:44c2::6...

Connected to uucp.eternal-september.org.
Escape character is '^]'.

login: ^]
telnet> quit
Connection closed.

Other hosts give me a TCP RST.

Most likely the stuff injected there will go to the reader innreport.

You have to ask the operator why that is the case.

[Marco Moock]

Am 28.12.2023 schrieb Paul Edwards <muta...@gmail.com>:

There are, see the innreport.

> No demand for bulk, efficient message transfer?

Maybe in de.comm.uucp or comp.mail.uucp are users that use it for that
purpose.
I remember there was also a discussion about a newer protocol that is a
bit similar to UUCP, but it intended to run over IP (I don't mean
UUCP over IP).

[Ray Banana]

Thus spake Paul Edwards <muta...@gmail.com>

> On 28/12/23 17:58, Paul Edwards wrote:
>> I tried:
>> telnet news.eternal-september.org 540

UUCP is running on uucp.eternal-september.org port 540,

> It occurs to me (I'm from Fidonet, not Usenet) -
> is this considered to be "peering" and thus not
> available to "normal" users?

That's it, peering, not reading. UUCP nodes use it to
download batches of articles from a peer server and import
them into their own servers.

Hence, UUCP nodes require a UUCP account (which is not the same as a
user account on the upstream serverand also need to have a peer defined
in the downstream server's newsfeeds file including the names of the
newsgroups they would like to receive. In addition, a batcher task on
the downstream server must be run at regular intervals and creates
batches sometimes referred to as bag files. The clients will then
connect to the downstream server, download the batch files and feed them
to a utility on their own machines that imports these batch files into
their own servers.

> In Fidonet end users could become "points" off
> an ordinary node, and that is what I was expecting
> UUCP to be the equivalent of.

[...]

> I must say - I wasn't expecting to come here and
> find no (ordinary) UUCP users at all. No demand
> for bulk, efficient message transfer?

Eternal-September currently has 5 UUCP users and 8200 NNTP users,
and that seems to be a representative ratio.

On the other hand, the server uucp.gnuu.de that shows up in
Eternal-September's peer list is actually a "normal" peer that offers
UUCP access to its clients. The pre-predecessor of Eternal-September
(banana.shacknet.nu) started off with uucp.gnuu.de as the sole (UUCP)
feed.

--
Пу́тін — хуйло́
https://www.eternal-september.org

[Ray Banana]

Thus spake Marco Moock <mm+s...@dorfdsl.de>

> Most likely the stuff injected there will go to the reader innreport.

Correct. As I mentioned in my previous post, UUCP clients require a peer
definition in the server's newsfeeds file and UUCP users can maintain
their newsgroup subscriptions via email. As the tool that handles these
emails directly modifies the server's newsfeeds file, I prefer to have
normal peers and UUCP clients in separate newsfeeds files and that means
different servers.

[Paul Edwards]

On 28/12/23 19:28, Ray Banana wrote:

> Thus spake Paul Edwards <muta...@gmail.com>
>
>> On 28/12/23 17:58, Paul Edwards wrote:
>>> I tried:
>>> telnet news.eternal-september.org 540
>
> UUCP is running on uucp.eternal-september.org port 540,

Yes - that's what I used originally (in my
original post). And I mentioned that my
userid and password caused a termination.

> On the other hand, the server uucp.gnuu.de that shows up in
> Eternal-September's peer list is actually a "normal" peer that offers
> UUCP access to its clients.

I went there, and google translated German to
English, and it looked to me like I could
register.

Do you happen to know:

1. Are non-Germans allowed to use this?

2. I didn't notice any fees listed.

3. Just to clarify - this is for ordinary
end users who wish to do UUCP, right? You
don't have to be one of these always-on
"peers", right?

Thanks. Paul.

[Ray Banana]

Thus spake Paul Edwards <muta...@gmail.com>

>> On the other hand, the server uucp.gnuu.de that shows up in
>> Eternal-September's peer list is actually a "normal" peer that offers
>> UUCP access to its clients.
> I went there, and google translated German to
> English, and it looked to me like I could
> register.
> Do you happen to know:
> 1. Are non-Germans allowed to use this?
> 2. I didn't notice any fees listed.

uucp.gnuu.de used to be a free (for their customers) service provided by
an ISP named germany.net that was acquired by Arcor which was eventually
acquired by Vodafone. (see https://en.wikipedia.org/wiki/Arcor_(telecommunications) ).

When Arcor discontinued the UUCP service along with the Usenet service,
the (former) news admin, Frank Kloeker a.k.a. Eumel, and a number of
users founded a registered association named GNUU e.V. that continued to
operate the service for members of the said association. The membership
fee is €50 for the first year and €35/year after that. In order to use
the service, you must first apply for membership, which will be approved
or denied by the president of the said association. Membership includes
participating in annual general meetings and so on. As I haven't been
involved in this for 14 or 15 years, you might want to contact
eumel at admin.gnuu.de for details.

hth

[Paul Edwards]

Sorry - am I reading that right?

You support "UUCP clients"? ie normal individuals
who only connect once/day to collect a mail bundle
or whatever?

If so, what is required to be one of those people?
I didn't see any mention on your website.

Thanks. Paul.

[Grant Taylor]

On 12/28/23 04:04, Paul Edwards wrote:
> In Fidonet end users could become "points" off an ordinary node,
> and that is what I was expecting UUCP to be the equivalent of.

I think that it's inaccurate to compare UUCP to points subtended off of
an ordinary node.

UUCP as a protocol is more akin to the communications used between any
two nodes, independent if of it's between two nodes or between a node
and a point.

N.B.

- I've not seen any things else have anything like the concept of
points subtended off of a node in the Internet. -- The closest I've
seen, not on the Internet, might be something in SNA which is nominally
a hierarchical network.

- Nor have I seen anything on the Internet use a hierarchical network
for machine to machine communications similar to how I understand FTN to
have originally worked. -- I don't know if contemporary uses of FTN
still strictly adhere to hierarchical routing; point -> local node ->
region / country node -> region / country node -> local node -> point.

Something else to keep in mind is that UUCP is / UUCP networks are a
store-and-forward mechanism where is much of the Internet is end-to-end
mechanism.

> But given that a "machine name" is part of the process, maybe they
> never had what Fidonet had, and maybe that is partly why Fidonet came
> up with something more appropriate/practical in the first place?

Based on my understanding, points are a special half way point between
an end user w/o a node and a full node. Something that a power user
would use but be dependent on the node they are subtended from in order
to route traffic for them.

> I must say - I wasn't expecting to come here and find no (ordinary)
> UUCP users at all. No demand for bulk, efficient message transfer?

UUCP is probably effectively unused by anyone other than hobbyists or
people with very specific use cases.

UUCP can be used as a transport mechanism for news (Usenet or
otherwise), email, file copy, remote command execution. But UUCP is
dependent on something providing a file and directions to send it to a
remote node. That means that UUCP is inherently push technology.

Conversely news readers are inherently pull technology.

As such, the pull nature of readers tend to be diametrically opposite to
push nature of UUCP.

N.B. there is a way for a UUCP system to probe a remote system to see if
there are any spooled files for it. As in my system can ""call your
system which gives your system an opportunity to push files it has
spooled to me over the same established connection. -- But this is
dependent on there being spooled, or pushed, articles on the remote
system. Which brings us back to the push nature of UUCP and pull nature
of news readers.

Peers feed, or push, all articles that match specific configured
patterns to to peers through some communications channel. That
communications channel can be NNTP or UUCP or something more esoteric.
But peers fundamentally push matching articles to each other.



--
Grant. . . .

[Jesse Rehmer]

It is on the homepage of https://www.eternal-september.org:

UUCP batches

news.eternal-september.org offers UUCP Batches on request.

Please send an email to peering (AT) eternal-september (DOT) org
news.eternal-september.org submits path information to the TOP1000 project
daily

[Grant Taylor]

On 12/28/23 09:19, Paul Edwards wrote:
> Sorry - am I reading that right?

I think you are mis-understanding.

> You support "UUCP clients"? ie normal individuals
> who only connect once/day to collect a mail bundle
> or whatever?

UUCP really only has peers.

Ray's news server has end users that run their own news server but are
peered with Ray's server via UUCP.

Ray's server feeds articles that match the peer's desired filters to the
end users via UUCP.

The process of feeding these articles means that the news server pushes
them into the UUCP queue to wait for the end user to connect and
retrieve the queued messages.

You might compare UUCP to a postal mailbox wherein you walk to your
local mailbox and check to see if there are any messages therein for
you. -- The postal carrier receives a bunch of messages and sorts them
appropriately. The messages that are for you get put into your mailbox.
Thus the postal carrier checks the larger number of UUCP peers
(everybody on your street, the post office down town, etc.).

You write a letter and push it into your mailbox waiting for the postal
carrier to come by and collect it. The carrier periodically checks your
mailbox to see if you have outgoing messages as well as depositing any
messages coming into you. -- This repeats at your local city level,
your regional level, your national level, etc on towards the recipient.
Much like the hierarchical nature of FTN.

You can't go to your mailbox and expect to choose a newsletter that
you've never had any association with. Instead you have to send an
outgoing message to ask that you be added to the mailing list.

> If so, what is required to be one of those people?

They will need to run their own news server. Or at least software that
can receive news via UUCP, ostensibly the rnews command. How true of a
news server it is is sort of independent as long as it can interact with
UUCP and receive articles.

> I didn't see any mention on your website.

People wanting UUCP is rare enough that it's now one of those things,
"ask if you want details" and is becoming the rare case by case type thing.

[Paul Edwards]

On 28/12/23 23:39, Grant Taylor wrote:
> On 12/28/23 04:04, Paul Edwards wrote:

> - Nor have I seen anything on the Internet use a hierarchical network
> for machine to machine communications similar to how I understand FTN to
> have originally worked. -- I don't know if contemporary uses of FTN
> still strictly adhere to hierarchical routing; point -> local node ->
> region / country node -> region / country node -> local node -> point.

That hierarchy is for email (and only nominally).

Any two ordinary nodes can choose to exchange
echoes (newsgroups) of any sort they choose.

And yes, points were basically for power users.

As UUCP should be.

No power users here?

Note that I wrote software to make the power user
software for Fidonet easier.

Now my plan is to do the same with Usenet. Gate
it (transparently) into my existing fidonet software.

I need to modify that software to make it transparent
though.

But it starts with UUCP bundles and I was surprised
it wasn't documented how to get those.

BFN. Paul.

[Paul Edwards]

1. I could have sworn that I had read something like
that years ago.

2. I swear I read every page multiple times looking
for it again recently and couldn't find it. Numerous
searches everywhere. No mention anywhere by anyone.

3. Thanks!!! I'll try my luck.

BFN. Paul.

[Adam H. Kerman]

Paul Edwards <muta...@gmail.com> wrote:

>>. . .

>I went there, and google translated German to
>English, and it looked to me like I could
>register.

>Do you happen to know:

>1. Are non-Germans allowed to use this?

>2. I didn't notice any fees listed.

>3. Just to clarify - this is for ordinary
>end users who wish to do UUCP, right? You
>don't have to be one of these always-on
>"peers", right?

That's what's wonderful about UUCP. It's the solution for a peer that's
not always on.

[Grant Taylor]

On 12/28/23 12:55, Adam H. Kerman wrote:
> That's what's wonderful about UUCP. It's the solution for a peer
> that's not always on.

UUCP is useful for more than that.

UUCP can also be used to bidirectionally transfer news / email / files
/ commands from dynamic locations too. E.g. UUCP-over-SSH initiated
from a road-warrior type client.

[Grant Taylor]

On 12/28/23 09:50, Paul Edwards wrote:
> As UUCP should be.

What's wrong with a power user running their own news / email / etc.
server using NNTP / SMTP / etc.?

That's how I started with Usenet 15+ years ago.

> No power users here?

I consider myself to be a power user and I don't route news / email
through UUCP.

> Now my plan is to do the same with Usenet. Gate it (transparently)
> into my existing fidonet software.

I'm not sure that you can /transparently/ gateway between networks. At
least not undetectable. Especially for people that view both sides.

> But it starts with UUCP bundles and I was surprised it wasn't
> documented how to get those.

Why do you say it starts with UUCP bundles?

I see some value in UUCP, but I don't see a need for it.

[Adam H. Kerman]

Absolutely.

[Thomas Hochstein]

Ray Banana wrote:

> When Arcor discontinued the UUCP service along with the Usenet service,
> the (former) news admin, Frank Kloeker a.k.a. Eumel, and a number of
> users founded a registered association named GNUU e.V. that continued to
> operate the service for members of the said association. The membership
> fee is €50 for the first year and €35/year after that. In order to use
> the service, you must first apply for membership, which will be approved
> or denied by the president of the said association. Membership includes
> participating in annual general meetings and so on.

Not necessarily; the general meetings, which are held online via matrix,
are usually attended by only one or two members in addition to the
president. :-)

> As I haven't been
> involved in this for 14 or 15 years, you might want to contact
> eumel at admin.gnuu.de for details.

Everything you wrote is still up-to-date.

-thh

[Paul Edwards]

On 29/12/23 03:53, Grant Taylor wrote:
> On 12/28/23 09:50, Paul Edwards wrote:
>> As UUCP should be.
>
> What's wrong with a power user running their own news / email / etc.
> server using NNTP / SMTP / etc.?

Ok, fair enough. That's what superpower users
are doing. I am after something different.

>> Now my plan is to do the same with Usenet. Gate it (transparently)
>> into my existing fidonet software.
>
> I'm not sure that you can /transparently/ gateway between networks. At
> least not undetectable. Especially for people that view both sides.

Maybe I used the wrong word.

I am only gating/converting it for my own use.

Those are the rules for end users of ES (so
it needs to be transparent - well - not fed
onwards).

However, if the internet dies for some reason,
I can still carry on communicating with anyone
I have an alternate way of reaching (modem or
walking to their house etc). THAT's what I want
to secure. But until that actually happens, for
Usenet to be the source of "echoes" (newsgroups).

>> But it starts with UUCP bundles and I was surprised it wasn't
>> documented how to get those.
>
> Why do you say it starts with UUCP bundles?

I wish to get efficient offline reading working.

Fidonetters (where I hung out) talked non-stop
about switching to UUCP. No-one talked about
running a news server.

BFN. Paul.

[Jon Ribbens]

There have been Fidonet/Usenet gateways for going on 40 years.
e.g. ufgate (which uses UUCP). But you don't need to use UUCP
anyway, you can use NNTP NEWNEWS to implement a pull feed if
you want, and you don't need special permission or configuration
from the news server operator to do so.

[VanguardLH]

Paul Edwards <muta...@gmail.com> wrote:

> Fidonetters (where I hung out) talked non-stop about switching to
> UUCP. No-one talked about running a news server.

There might be other leeching NNTP clients that behave as local NNTP
server (act as an local NNTP client to retrieve all articles from
specified newsgroups) for personal use, but the ones I remember hearing
about are:

Hamster [Next] (ask for download sources in hamster.* newsgroups)
Leafnode (https://www.leafnode.org/)

Those look like NNTP clients to an NNTP server. They aren't peering
with the NNTP server, especially since the users do not have peering
relationships with the server, but just clients yanking down articles
from the server just like how an NNTP client operates. While they can
be accessed by any NNTP client, they are usually setup only for personal
use, like someone that wants to have their newsreader connect to a
server that is local, or a company that wants to yank newsgroups
relevant to their employees.

[Paul Edwards]

On 29/12/23 10:03, Jon Ribbens wrote:
> On 2023-12-29, Paul Edwards <muta...@gmail.com> wrote:

>> Fidonetters (where I hung out) talked non-stop
>> about switching to UUCP. No-one talked about
>> running a news server.
>
> There have been Fidonet/Usenet gateways for going on 40 years.
> e.g. ufgate (which uses UUCP).

Sure. And I was never involved in that. And now
I am getting involved in it. On my own terms.

> But you don't need to use UUCP
> anyway, you can use NNTP NEWNEWS to implement a pull feed if
> you want, and you don't need special permission or configuration
> from the news server operator to do so.

Sure - the non-power-user way of doing things.

Same as on a BBS you could use Bluewave (I think)
to read messages offline.

But power users contacted the sysop and asked to
become a point.

I made my BBS points-only, effectively. There was
minimal software and messenging available to set
up a point.

Note that one of the reasons I was late to the
party is that one of the components I wished to
replace (before getting to UUCP) was the OS.

That took about 35 years (pdos.org).

BFN. Paul.

[Jon Ribbens]

On 2023-12-29, Paul Edwards <muta...@gmail.com> wrote:
> On 29/12/23 10:03, Jon Ribbens wrote:
>> But you don't need to use UUCP
>> anyway, you can use NNTP NEWNEWS to implement a pull feed if
>> you want, and you don't need special permission or configuration
>> from the news server operator to do so.
>
> Sure - the non-power-user way of doing things.

It really makes little difference if you're intermittently connected.
It's not like point vs node.

[Paul Edwards]

This could be a timed connection.

The goal is to have the data all compressed
and waiting for you when you hit it.

Even if you're hitting it with a Commodore 64.

BFN. Paul.

[Grant Taylor]

On 12/28/23 19:53, Paul Edwards wrote:
> I am only gating/converting it for my own use.

Fair enough.

Much of my concern around the gatewaying is around header differences
and loop avoidance (multiple gateways)

> However, if the internet dies for some reason, I can still carry on
> communicating with anyone I have an alternate way of reaching (modem
> or walking to their house etc). THAT's what I want to secure. But
> until that actually happens, for Usenet to be the source of "echoes"
> (newsgroups).

ACK

Please elaborate on your choice of the word "secure". Do you mean
something like "acquire" / "procure" or do you mean something like
"privacy"? The former is a lot easier than the latter.

> I wish to get efficient offline reading working.

ACK

> Fidonetters (where I hung out) talked non-stop about switching to
> UUCP. No-one talked about running a news server.

I've spent some time thinking about this over the years -- I've not yet
been motivated to do more than think or talk about it. If I were to do
this, I would very likely set up a local NNTP (and SMTP) server that had
a UUCP connection to counterparts on the Internet. Each end would batch
things to go to the other end and then I'd periodically cause
connections to be made do a batch transfer.

N.B. the Internet end would probably not initiate the connection and
instead rely on the private end to initiate the connection and then send
the inbound traffic over the same outbound connection that the local
server initiated.

[Grant Taylor]

On 12/28/23 22:47, Paul Edwards wrote:
> This could be a timed connection.

ACK

> The goal is to have the data all compressed
> and waiting for you when you hit it.

I don't know about /compression/ per se, but I do know that UUCP could
definitely handle the queued transmission.

> Even if you're hitting it with a Commodore 64.

I don't think I'd want to mess with decompression on a Commodore 64.
But you do you. ;-)

[Paul Edwards]

On 29/12/23 14:02, Grant Taylor wrote:
> On 12/28/23 19:53, Paul Edwards wrote:

BTW - I am surprised that I ended up having
this conversation here. I was expecting to
have it elsewhere in due course. But this
could be my new home. :-)

>> The goal is to have the data all compressed
>> and waiting for you when you hit it.

> I don't know about /compression/ per se, but I
> do know that UUCP could definitely handle the
> queued transmission.

UUCP traditionally compressed news, but didn't
compress email. I don't know why.

But I only care about news.

>> Even if you're hitting it with a Commodore 64.

> I don't think I'd want to mess with decompression
> on a Commodore 64. But you do you.

If you are phoning the other side of the world
in the 1980s, running decompression overnight is
not an issue.

Sorry about the unstated assumption - reliance
on modern computers or whatever is something I
consider "cheating" and/or "not secure" if for
some reason you no longer have access to them.

>> However, if the internet dies for some reason, I can still carry on
>> communicating with anyone I have an alternate way of reaching (modem
>> or walking to their house etc). THAT's what I want to secure. But
>> until that actually happens, for Usenet to be the source of "echoes"
>> (newsgroups).
>
> ACK
>
> Please elaborate on your choice of the word "secure". Do you mean
> something like "acquire" / "procure" or do you mean something like
> "privacy"? The former is a lot easier than the latter.

Neither. :-)

Ask the President to "secure" the southern border. :-)

(I'm not American BTW).

Have you seen the movie "Idiocracy"?

So - can you construct the technology you depend on?

I can construct most of the software side. Not
even construct it - I have the source code. It's
in a language I understand (C90). The stuff I am
so far responsible for (the OS) is as simple as
it can be, and thus understandable (especially
the flagship PDOS-generic).

It's all public domain (not one of the freeware
licenses). It's using technology more than 20
years old so free of patent.

I haven't yet burned all this onto a commercially
pressed CD, but I have made the effort to find out
if the pits are visible via a microscope (yes, but
difficult).

I also haven't organized plastic punch cards.

I don't trust magic 2 TB hard disks as far as
I can throw them. Well. As far as I can throw
a washing machine.

"Steamboat Willie" is coming out of copyright
in around 3 days. I don't have the ability to
decompress MP4, but even if I did, I don't trust
it. I want to see a text file (see artwork.txt
in c:\baby on PDOS/386 - "showpic" to display).

I want to know how to get that onto film, and
I want it on film, as a backup.

In "Mad Max - Beyond Thunderdome" they were left
with one of those slide viewers. That's it.

Will Putin start a nuclear war, and will he target
uninvolved major cities like Manila (so that they
don't benefit post-war)? I currently live elsewhere
in the Philippines, so I won't be touched, but will
I need to get new CPUs from universities in the
Philippines rather than overseas because overseas
trade ceased to be a thing?

>> Fidonetters (where I hung out) talked non-stop about switching to
>> UUCP. No-one talked about running a news server.
>
> I've spent some time thinking about this over the years -- I've not yet
> been motivated to do more than think or talk about it. If I were to do
> this, I would very likely set up a local NNTP (and SMTP) server that had
> a UUCP connection to counterparts on the Internet.

That sounds like it requires multitasking. Got
a public domain multitasking OS available? Mine
is single-tasking (quibbling aside). Will you
ever be restricted to only being allowed to use
public domain software? I don't know what
circumstances that will occur in, but what I do
know is that the vast majority of freeware authors
are very reluctant (as in - refuse point blank),
to release their software to the public domain.

What if Fujitsu employs you to do a clean room
implementation of news reading and insists that
you can only use public domain software in a
literal room so that they can be squeaky clean?

The majority of the world believes in some sort
of god(s). What if he/she/they/it send you back
in time because the current universe has reached
some sort of dead end? Are you ready? That
happened in Terminator. The robot arrived naked.
That could be you. What would you do?

If I'm taken back to the year 1000, it's a moot
point - the hardware won't exist.

But I want to cover the years where the hardware
exists but not the software. Will I replace CRLF
with just LF or have a separate (or sole) NL? I
don't yet have the understanding to answer that.
EBCDIC has all of CR, LF and NL.

Several hours ago I was attempting to determine
whether EBCDIC ANSI (X3.64) terminals could be
made to work over an OSA-ICC TN3270 link.

I don't have an answer yet. The person testing
it went to sleep. Is there a show-stopper? I
don't know. I don't have a sufficient understand.
I haven't secured that concept of ANSI X3.64
being universally possible. Well - I sort of have.
It can always be put onto a proper 3270 data
stream (which I've already demonstrated). But can
it go across the link raw? Inquiring minds would
like to know.

Everyone is going to be "wise" (of course it will/
won't work - don't you have a basic understanding
of xyz?) after the event. But I'm on this side
of the event. I haven't noticed the mainframers
I hang out with making bold predictions. Some of
them got burned on previous predictions (S/380).

BFN. Paul.

[Ray Banana]

Thus spake Paul Edwards <muta...@gmail.com>

>> news.eternal-september.org offers UUCP Batches on request.
>> Please send an email to peering (AT) eternal-september (DOT) org
>> news.eternal-september.org submits path information to the TOP1000 project
>> daily
> 1. I could have sworn that I had read something like
> that years ago.

> 2. I swear I read every page multiple times looking
> for it again recently and couldn't find it. Numerous
> searches everywhere. No mention anywhere by anyone.
> 3. Thanks!!! I'll try my luck.

Hi,

since there seems to be some interest in UUCP in this group, I'm
replying here to clarify what UUCP support Eternal-September offers and
what prerequisites you need to meet in order to use it.

Eternal-September offers UUCP accounts to its users that allow them to
connect to the UUCP service, subscribe to newsgroups and download
batches of articles to their machines. These batches contain compressed
articles along with a call to a utility to process the batch,

The user needs the UUCP software installed and configured on their local
machines. If your local machine is not running Unix or Linux, I will not
be able to provide support for your local installation.

There are UUCP packages for Windows, MSDOS and OS/2 available at

https://www.uupc.net/pub/uupc/prod/

but these seem to be rather old (2002). The source code is also
available on the site.

If you do not have a newsserver installed on your local machine, you
will need to write your own tool for processing the batch files in
whatever way you intend to store the downloaded articles.

You will also need a newsreader or mail client that can read the mbox
format in order to properly render the articles (encoding, character
set, mime format etc.).

HTH

[Paul Edwards]

On 29/12/23 16:52, Ray Banana wrote:

> since there seems to be some interest in UUCP in this group, I'm
> replying here to clarify what UUCP support Eternal-September offers and
> what prerequisites you need to meet in order to use it.

Thanks.

> The user needs the UUCP software installed and configured on their local
> machines. If your local machine is not running Unix or Linux, I will not
> be able to provide support for your local installation.

Understood. :-)

> There are UUCP packages for Windows, MSDOS and OS/2 available at
>
> https://www.uupc.net/pub/uupc/prod/
>
> but these seem to be rather old (2002). The source code is also
> available on the site.

I do my development on Windows 2000 (last version
that doesn't need "verification" or whatever they
call it under Virtualbox under Kylin Linux
and I have Visual Studio 2005 Professional, so I got
the source code and opened a 32-bit command prompt
and did:

D:\devel\uupc>nmake -f nmake.mak

and after a lot of weird warnings (it considers
strcpy obsolete?), it completed successfully.

uucico crashes deliberately (I checked the
source code line reported) if there is any
configuration error, so I did this:

set UUPCSYSRC=fred.cfg

and created a dummy config file, but it needs
more than just a dummy.

So assuming I am given an account I'll try to
find out what it wants.

> If you do not have a newsserver installed on your local machine, you
> will need to write your own tool for processing the batch files in
> whatever way you intend to store the downloaded articles.

Sure always the intention.

> You will also need a newsreader or mail client that can read the mbox
> format in order to properly render the articles (encoding, character
> set, mime format etc.).

Sure. Part of the challenge.

BFN. Paul.

[Grant Taylor]

On 12/29/23 01:10, Paul Edwards wrote:
> BTW - I am surprised that I ended up having this conversation here. I
> was expecting to have it elsewhere in due course. But this could be
> my new home. :-)

I've had lots of interesting conversations in unexpected places.

> UUCP traditionally compressed news, but didn't compress email. I
> don't know why.

I sort of suspect it's not UUCP doing the compression. Rather it's
whatever the source of the data is handing pre-compressed data off to UUCP.

I sort of suspect that similar could be done with mail.

> But I only care about news.
>

> If you are phoning the other side of the world in the 1980s, running
> decompression overnight is not an issue.

I suppose not.

> Sorry about the unstated assumption - reliance on modern computers
> or whatever is something I consider "cheating" and/or "not secure"
> if for some reason you no longer have access to them.

I think there are a lot more capable systems that are far from modern
computers. Anything built in the '90s or even early '00s qualifies and
is still more than 20 years old.

> Neither. :-)

I haven't detected a language barrier with you before Paul, but I don't
know what your native language is.

As it is, I'm still not sure what you mean by "secure" in the context of
what you want to secure.

> I can construct most of the software side. Not even construct it -
> I have the source code. It's in a language I understand (C90). The
> stuff I am so far responsible for (the OS) is as simple as it can be,
> and thus understandable (especially the flagship PDOS-generic).

I'm not sure if, much less how, this relates to securing something.

> I want to know how to get that onto film, and I want it on film,
> as a backup.

I think you're going off into a different direction.

> That sounds like it requires multitasking.

I don't think so.

> What if Fujitsu employs you to do a clean room implementation of news
> reading and insists that you can only use public domain software in
> a literal room so that they can be squeaky clean?

I wouldn't consider that to be a problem.

News, as in NNTP, is text. It's a matter of handling text in well
defined, and fairly easily discernible formats.

> Several hours ago I was attempting to determine whether EBCDIC ANSI
> (X3.64) terminals could be made to work over an OSA-ICC TN3270 link.

Seeing how OSA is the defacto network interface to IBM mainframes for
the last (almost) two decades and they are supporting emulated and
physical terminals, I'm quite certain that they can.

[Paul Edwards]

On 30/12/23 06:49, Grant Taylor wrote:
> On 12/29/23 01:10, Paul Edwards wrote:

>> UUCP traditionally compressed news, but didn't compress email. I don't
>> know why.
>
> I sort of suspect it's not UUCP doing the compression. Rather it's
> whatever the source of the data is handing pre-compressed data off to UUCP.
>
> I sort of suspect that similar could be done with mail.

I saw somewhere recently that (some?) UUCP
systems now compress mail. I'm just saying
that traditionally they didn't.

>> Neither. :-)
>
> I haven't detected a language barrier with you before Paul, but I don't
> know what your native language is.

English (Australian).

> As it is, I'm still not sure what you mean by "secure" in the context of
> what you want to secure.

https://www.dictionary.com/browse/secure

free from or not exposed to danger or harm; safe.

(first meaning)

I mentioned getting my source code on punch
cards. Those will survive an EMP strike. As
just one example.

I remember reading about a patented drug that
was sold to a new person who jacked the price
up 50 or 70 fold overnight. If you were dependent
on that drug, did you have a contingency plan for
that?

What's your backup plan if you're Russian and
can no longer obtain either x86 chips or Windows
due to sanctions?

Lots of such considerations.

>> I can construct most of the software side. Not even construct it - I
>> have the source code. It's in a language I understand (C90). The stuff
>> I am so far responsible for (the OS) is as simple as it can be, and
>> thus understandable (especially the flagship PDOS-generic).
>
> I'm not sure if, much less how, this relates to securing something.

I want to be able to exchange public newsgroups/echomail,
even in the event of the internet going down. Even in the
event of Intel and AMD conspiring to put a drop dead date
in their CPUs set to 2024-01-01. I bought a Chinese-made
computer with a Zhaoxin chip so I'll be last man standing.

I want to secure the technology (newsgroup exchange) that
I use. People who know how to maintain it. People who have
the source code. Punch card backup to survive EMP. Punch
cards as durable plastic to survive a 1000 year gap in
human development thanks to a virus that wipes out 99.9%
of the world population right after a global nuclear war
with deliberate targeting of any industrialized city to
ensure that no-one gets an advantage post-war.

> News, as in NNTP, is text. It's a matter of handling text in well
> defined, and fairly easily discernible formats.

Absolutely. But does the TCP/IP stack you use
to get that depend on multitasking and virtual
memory to be available? What if the 8088 was
the last chip that Intel didn't secretly put
a drop dead date of 2024-01-01 into? I have
bought 2 Book 8088 computers. I'd like to
continue talking to you personally when the
only computers still working are 8088-based.
I'm happy to loan (not give - mine remaining
one might break so I need it back) my second
Book 8088. Does your NNTP reader have an
8088 target? If it was written in C90 it
shouldn't be a problem. Was it? Have you
ever looked?

>> Several hours ago I was attempting to determine whether EBCDIC ANSI
>> (X3.64) terminals could be made to work over an OSA-ICC TN3270 link.
>
> Seeing how OSA is the defacto network interface to IBM mainframes for
> the last (almost) two decades and they are supporting emulated and
> physical terminals, I'm quite certain that they can.

How can you be certain that some component of the
3270 link doesn't insist on a valid 3270 data stream?

I'm trying to send EBCDIC ANSI over it which is
invalid.

However, it's very close to having the basics
working now. z/PDOS running under z/VM is able
to IPL and display all the startup messages.

And even if something is found that "validates"
a 3270 data stream, I suspect that a valid 3270
data stream can be constructed that will get past
that check, without forcing the entire thing to
be transferred as hex text.

But no, that wasn't intuitively obvious to me.
As far as I know, people have consistently said
that it is impossible to support character-based
terminals on a mainframe - the supporting
infrastructure isn't there - everything is block
mode.

Even as I inched closer - mainframers only said
that it might work over the OSA-ICC, not that it
definitely will.

BFN. Paul.

[Grant Taylor]

On 12/29/23 20:53, Paul Edwards wrote:
> I saw somewhere recently that (some?) UUCP systems now compress
> mail. I'm just saying that traditionally they didn't.

Fair enough.

N.B. what has been done in the past does not dictate was can be done in
the future.

> English (Australian).

ACK

> https://www.dictionary.com/browse/secure
>
> free from or not exposed to danger or harm; safe.

Okay. That's decidedly different than I see "secure" in most computer /
technology contexts. But I can appreciate that definition.

> I mentioned getting my source code on punch cards. Those will survive
> an EMP strike. As just one example.

Okay.

But will punch cards survive a fire or flood?

What about three decks of punch cards being mixed together? What if the
cards aren't marked as clearly as one would hope?

Point being that each thing has it's strengths and weaknesses.

> I remember reading about a patented drug that was sold to a new person
> who jacked the price up 50 or 70 fold overnight. If you were dependent
> on that drug, did you have a contingency plan for that?

Sometimes there isn't a contingency plan.

> What's your backup plan if you're Russian and can no longer obtain
> either x86 chips or Windows due to sanctions?

I thought that there was an x86 compatible chip manufactured in the
Russian Federation at some time in the past. Maybe not.

Also, I suspect that they already have copies of Windows that can be
installed on new systems. ;-)

> Lots of such considerations.

Yes.

> I want to be able to exchange public newsgroups/echomail, even in the
> event of the internet going down. Even in the event of Intel and AMD
> conspiring to put a drop dead date in their CPUs set to 2024-01-01. I
> bought a Chinese-made computer with a Zhaoxin chip so I'll be last
> man standing.

All Intel and AMD CPUs bricking at a date wouldn't end things as we know
them. They might seriously limit them. But save for physical self
destruction, change the date to something in the past.

Or as you mentioned, CPU from someone other than Intel and AMD.

> I want to secure the technology (newsgroup exchange) that I use. People
> who know how to maintain it. People who have the source code. Punch
> card backup to survive EMP. Punch cards as durable plastic to survive
> a 1000 year gap in human development thanks to a virus that wipes
> out 99.9% of the world population right after a global nuclear war
> with deliberate targeting of any industrialized city to ensure that
> no-one gets an advantage post-war.

LOL

Just because the raw material might survive doesn't mean that anyone
will know how to use it.

I've been told by a friend that I trust that the U.S.A. has things quite
well documented on what needs to be made, as in blue prints, but we
can't make it because the method to make them to the specs was common
knowledge at the time. Now it's so uncommon that it's extremely rare to
the point that things can't be made. NASA and things from the '60s
comes to mind.

> Absolutely. But does the TCP/IP stack you use to get that depend on
> multitasking and virtual memory to be available?

There are multiple TCP/IP stacks for MS-DOS which is ostensibly
single-tasking and doesn't have native virtual memory.

> What if the 8088 was the last chip that Intel didn't secretly
> put a drop dead date of 2024-01-01 into? I have bought 2 Book 8088
> computers. I'd like to continue talking to you personally when the only
> computers still working are 8088-based. I'm happy to loan (not give
> - mine remaining one might break so I need it back) my second Book
> 8088. Does your NNTP reader have an 8088 target? If it was written
> in C90 it shouldn't be a problem. Was it? Have you ever looked?

The news reader that I'm typing this in probably does not.

But there is a text editor and a telnet client for the aforementioned
TCP/IP stack for MS-DOS.

> But no, that wasn't intuitively obvious to me. As far as I
> know, people have consistently said that it is impossible to
> support character-based terminals on a mainframe - the supporting
> infrastructure isn't there - everything is block mode.

I sort of suspect that part of that is OS centric. As In I think there
are other OSs for the mainframe that aren't 3270 and / or EBCDIC. I
/think/.

> Even as I inched closer - mainframers only said that it might work
> over the OSA-ICC, not that it definitely will.

As I understand it, OSA's ICC is a complex server gateway platform
integral to the OSA card itself. There have been other things from IBM
over the years that do what the OSA ICC does. Some of them are what
convert between ASCII over serial lines and EBCDIC 3270 data streams
into the mainframe OS.

[Paul Edwards]

On 30/12/23 13:45, Grant Taylor wrote:
> On 12/29/23 20:53, Paul Edwards wrote:
>> I saw somewhere recently that (some?) UUCP systems now compress mail.
>> I'm just saying that traditionally they didn't.
>
> Fair enough.
>
> N.B. what has been done in the past does not dictate was can be done in
> the future.

Sure.

>> I mentioned getting my source code on punch cards. Those will survive
>> an EMP strike. As just one example.
>
> Okay.
>
> But will punch cards survive a fire or flood?
>
> What about three decks of punch cards being mixed together? What if the
> cards aren't marked as clearly as one would hope?
>
> Point being that each thing has it's strengths and weaknesses.

Sure. I'm not going to put the sole copy of
my source code onto punch cards.

>> I remember reading about a patented drug that was sold to a new person
>> who jacked the price up 50 or 70 fold overnight. If you were dependent
>> on that drug, did you have a contingency plan for that?
>
> Sometimes there isn't a contingency plan.

Right. So that's my interest. Making sure there
is one. Like someone organizing a backup of these
newsgroups to replace Google. And yes, there are
people already. But I'd like to be one of them.

>> What's your backup plan if you're Russian and can no longer obtain
>> either x86 chips or Windows due to sanctions?
>
> I thought that there was an x86 compatible chip manufactured in the
> Russian Federation at some time in the past. Maybe not.
>
> Also, I suspect that they already have copies of Windows that can be
> installed on new systems. ;-)

Ok, but if they need bug fixes for that, what
are they going to do? They don't have the source
code. So. They could use Linux instead. But then
after any effort they have made, they won't be
able to make a profit post-sanctions by
close-sourcing their software like Microsoft feels
the need to do.

So they're sort of stuck (by one definition).

So that's why I have spent 30 years (not full-time)
writing PDOS.

> All Intel and AMD CPUs bricking at a date wouldn't end things as we know
> them. They might seriously limit them. But save for physical self
> destruction, change the date to something in the past.

It is the physical self-destruction I am thinking
of. Or needing a code to unlock.

> Or as you mentioned, CPU from someone other than Intel and AMD.

A bit late by then.

>> I want to secure the technology (newsgroup exchange) that I use.
>> People who know how to maintain it. People who have the source code.
>> Punch card backup to survive EMP. Punch cards as durable plastic to
>> survive a 1000 year gap in human development thanks to a virus that
>> wipes out 99.9% of the world population right after a global nuclear
>> war with deliberate targeting of any industrialized city to ensure
>> that no-one gets an advantage post-war.
>
> LOL
>
> Just because the raw material might survive doesn't mean that anyone
> will know how to use it.

Sure. So something else to take into account.

> I've been told by a friend that I trust that the U.S.A. has things quite
> well documented on what needs to be made, as in blue prints, but we
> can't make it because the method to make them to the specs was common
> knowledge at the time. Now it's so uncommon that it's extremely rare to
> the point that things can't be made. NASA and things from the '60s
> comes to mind.

Ok. And even if Americans can make something,
if the Philippines is under sanctions for some
reason, can the Philippines independently make it?

>> Absolutely. But does the TCP/IP stack you use to get that depend on
>> multitasking and virtual memory to be available?
>
> There are multiple TCP/IP stacks for MS-DOS which is ostensibly
> single-tasking and doesn't have native virtual memory.

Ok. I'll probably need to get to that (with
PDOS/386) one day.

>> What if the 8088 was the last chip that Intel didn't secretly put a
>> drop dead date of 2024-01-01 into? I have bought 2 Book 8088
>> computers. I'd like to continue talking to you personally when the
>> only computers still working are 8088-based. I'm happy to loan (not
>> give - mine remaining one might break so I need it back) my second
>> Book 8088. Does your NNTP reader have an 8088 target? If it was
>> written in C90 it shouldn't be a problem. Was it? Have you ever looked?
>
> The news reader that I'm typing this in probably does not.
>
> But there is a text editor and a telnet client for the aforementioned
> TCP/IP stack for MS-DOS.

Sure. And even though I use Windows 2000 to do
my development, I stay at the command prompt
and am aware of each command I type and whether
that command/program exists in PDOS/386 yet, so
that I have a backstop.

>> But no, that wasn't intuitively obvious to me. As far as I know,
>> people have consistently said that it is impossible to support
>> character-based terminals on a mainframe - the supporting
>> infrastructure isn't there - everything is block mode.
>
> I sort of suspect that part of that is OS centric. As In I think there
> are other OSs for the mainframe that aren't 3270 and / or EBCDIC. I
> /think/.

Yes, z/Linux meets that description.

But I am after EBCDIC, but not 3270.

>> Even as I inched closer - mainframers only said that it might work
>> over the OSA-ICC, not that it definitely will.
>
> As I understand it, OSA's ICC is a complex server gateway platform
> integral to the OSA card itself. There have been other things from IBM
> over the years that do what the OSA ICC does. Some of them are what
> convert between ASCII over serial lines and EBCDIC 3270 data streams
> into the mainframe OS.

Sure. But I don't want 3270. I want to be able
to create fullscreen applications in a portable
manner.

And I chose ANSI X3.64 for that (because ANSI
already had), and wondered what the barrier to
having that on the mainframe was.

I was surprised 35 years ago when I didn't see
a version of MSDOS on the mainframe. And I
wondered why. Even for restricted use.

BFN. Paul.

[Grant Taylor]

On 12/30/23 16:29, Jukka Lahtinen wrote:
> Back in the 1990's one Fidonet sysop asked me if I'd like to become
> a private node, polling him. I did that a while, then private nodes
> became discouraged and he asked me to change to a point, which I did.

I don't think I've ever heard of a private node before.

I'll have to look that up.

[Grant Taylor]

On 12/30/23 00:21, Paul Edwards wrote:
> Ok, but if they need bug fixes for that, what are they going
> to do? They don't have the source code.

Source code is not strictly /required/ to patch software. Source code
just makes it multiple orders of magnitude simpler.

Take a look at companies releasing micro patches for Microsoft products
before Microsoft releases official patches.

> So they're sort of stuck (by one definition).

Not necessarily.

> Ok. And even if Americans can make something, if the Philippines is
> under sanctions for some reason, can the Philippines independently
> make it?

Capability and technical acumen are completely independent of legality.
Russia has had the former for many decades and did make / clone things
without the latter.

> Sure. But I don't want 3270. I want to be able to create fullscreen
> applications in a portable manner.

Mainframe has LONG had full screen applications. One of the most
prominent is ISPF.

> I was surprised 35 years ago when I didn't see a version of MSDOS on
> the mainframe. And I wondered why. Even for restricted use.

IBM released DOS for the mainframe. It's just not PC-DOS or any of it's
other names.

MS-DOS et al. is targeting a completely different function than IBM's
DOS on the mainframe. VASTLY different hardware and user bases.

[Paul Edwards]

On 31/12/23 09:52, Grant Taylor wrote:

>> Ok. And even if Americans can make something, if the Philippines is
>> under sanctions for some reason, can the Philippines independently
>> make it?
>
> Capability and technical acumen are completely independent of legality.
> Russia has had the former for many decades and did make / clone things
> without the latter.

The trouble is if you haven't done it legally then
you won't be able to market it in the west, so the
development effort is wasted.

That's my interest in these rogue nations. I expect
them to eventually become non-rogue, and then I want
that development effort to be sellable.

>> Sure. But I don't want 3270. I want to be able to create fullscreen
>> applications in a portable manner.
>
> Mainframe has LONG had full screen applications. One of the most
> prominent is ISPF.

I said "portable". ANSI X3.64 is the standard,
not 3270. So I'm looking to find out why people
aren't following the standard.

>> I was surprised 35 years ago when I didn't see a version of MSDOS on
>> the mainframe. And I wondered why. Even for restricted use.
>
> IBM released DOS for the mainframe. It's just not PC-DOS or any of it's
> other names.

DOS/VS is just a name. There's no point mentioning
it just because the names are coincidental. They
don't even own that anymore anyway.

Their main product has always been MVS (now z/OS).

> MS-DOS et al. is targeting a completely different function than IBM's
> DOS on the mainframe. VASTLY different hardware and user bases.

C90-compliant programs work just fine on IBM hardware.

And less than 24 hours ago I got EBCDIC ANSI terminals
to work on that IBM hardware.

So now when running z/PDOS you can't actually tell
whether you're on a mainframe or not (quibbling aside).

Which is exactly what I hoped to achieve.

BFN. Paul.

[Grant Taylor]

On 12/31/23 06:30, Jukka Lahtinen wrote:
> It was technically a Fidonet node, but for private use, not open
> for callers. I had one phone line and didn't want to dedicate it to
> keeping a bbs online.

Oh!

That makes more sense.

Thank you for clarifying Jukka. :-)

[Grant Taylor]

On 12/31/23 07:50, Paul Edwards wrote:
> The trouble is if you haven't done it legally then you won't be able
> to market it in the west, so the development effort is wasted.

Either I'm not understanding you -or- you keep moving the goal post.

The companies that have the micro-patches that I'm talking about are in
business selling them for profit.

> I said "portable". ANSI X3.64 is the standard, not 3270. So I'm
> looking to find out why people aren't following the standard.

The wonderful thing about standards is that we have so many to choose from.

3270 /is/ a standard. Just not the standard that you want to use.

> DOS/VS is just a name. There's no point mentioning it just because
> the names are coincidental. They don't even own that anymore anyway.

I question the veracity that last statement.

> Their main product has always been MVS (now z/OS).

For something to be the /main/ product implies there is at least one
other product.

> C90-compliant programs work just fine on IBM hardware.
>
> And less than 24 hours ago I got EBCDIC ANSI terminals to work on
> that IBM hardware.
>
> So now when running z/PDOS you can't actually tell whether you're on
> a mainframe or not (quibbling aside).
>
> Which is exactly what I hoped to achieve.

Good for you.

[Daniel65]

Grant Taylor wrote on 1/1/24 4:11 am:

Sounds like a regular "Dial-Up" account to me! Dial in, do what you want
to do (get mail, upload my news posts, download other news posts, do
whatever browsing) then get off-line to read your mail/news.

Do it all again, maybe. (Wash/Rinse/Repeat)
--
Daniel

[Eric Pozharski]

with <umrriq$1oalv$1...@dont-email.me> Paul Edwards wrote:
> On 31/12/23 09:52, Grant Taylor wrote:

>>> Ok. And even if Americans can make something, if the Philippines is
>>> under sanctions for some reason, can the Philippines independently
>>> make it?
>>
>> Capability and technical acumen are completely independent of
>> legality. Russia has had the former for many decades and did make /
>> clone things without the latter.
>
> The trouble is if you haven't done it legally then you won't be able
> to market it in the west, so the development effort is wasted.
>
> That's my interest in these rogue nations. I expect them to eventually
> become non-rogue, and then I want that development effort to be
> sellable.

Excuse me, what is it you're smoking? I'm a bit scared, what if it's
contageous?

Rogue nations don't work this way. Like at all. Let me introduce one
quote (translation is mine and non-canonical). This quote is
attributed and not in collected works. The person who said that was
deeply involved in creation of first rogue nation. Name is -- Lenin
(are you aware of his works?). So here it comes:

Slogan 'catch up and overtake America' first of all means
corrupting by all means necessary their economical and political
balance, their will to fight and resist. Only then it is
possible for us to catch up and overtake.

Your ideas of how rogue nations function are probably taken out of ass.

*CUT* [ 36 lines 3 levels deep]

--
Torvalds' goal for Linux is very simple: World Domination
Stallman's goal for GNU is even simpler: Freedom

[🌈💐🌻🌺🌹🌻💐🌷🌺🌈Jen🌈💐🌻🌺🌹🌻💐🌷🌺🌈 Dershmender 💐🌻🌺🌹🌻💐🌷🌺🐶笛🌈💐🌻🌺🌹🌻💐🌷🌺🌈]

On Mon, 01 Jan 2024 10:14:48 +0000, LO AND BEHOLD; Eric Pozharski
<why...@pozharski.name> determined that the following was of not great
importance to Eric Pozharski <why...@pozharski.name> and subsequently
decided to NOT freely share it with us in
<slrnup544o...@freight.zombinet>:

=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= with <umrriq$1oalv$1...@dont-email.me> Paul Edwards wrote:
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= On 31/12/23 09:52, Grant Taylor wrote:
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Ok. And even if Americans can make something, if the Philippines is
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= under sanctions for some reason, can the Philippines independently make
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= it?
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Capability and technical acumen are completely independent of legality.
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Russia has had the former for many decades and did make / clone things
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= without the latter.
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= The trouble is if you haven't done it legally then you won't be able to
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= market it in the west, so the development effort is wasted. That's my
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= interest in these rogue nations. I expect them to eventually become
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= =?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= non-rogue, and then I want that development effort to be sellable.
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Excuse me, what is it you're smoking? I'm a bit scared, what if it's
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= contageous?
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Rogue nations don't work this way. Like at all. Let me introduce one
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= quote (translation is mine and non-canonical). This quote is
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= attributed and not in collected works. The person who said that was
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= deeply involved in creation of first rogue nation. Name is -- Lenin
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= (are you aware of his works?). So here it comes:
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Slogan 'catch up and overtake America' first of all means corrupting
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= by all means necessary their economical and political balance, their
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= will to fight and resist. Only then it is possible for us to catch up
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= and overtake.
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= Your ideas of how rogue nations function are probably taken out of ass.
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?= *CUT* [ 36 lines 3 levels deep]
=?UTF-8?B?8J+Ps++4j+KAjfCfjIg=?=

Is this the thread for people who have been banned by Ray Bannanner for not following his Romper Room Rulez?

--

Golden Killfile, June 2005
KOTM, November 2006
Bob Allisat Memorial Hook, Line & Sinker, November 2006
Special Ops Cody Memorial Purple Heart, November 2006
Special Ops Cody Memorial Purple Heart, September 2007
Tony Sidaway Memorial "Drama Queen" Award, November 2006
Busted Urinal Award, April 2007
Order of the Holey Sockpuppet, September 2007
Barbara Woodhouse Memorial Dog Whistle, September 2006
Barbara Woodhouse Memorial Dog Whistle, April 2008
Tinfoil Sombrero, February 2007
AUK Mascot, September 2007
Putting the Awards Out of Order to Screw With the OCD Fuckheads, March 2016

[Grant Taylor]

On 1/1/24 02:27, Daniel65 wrote:
> Sounds like a regular "Dial-Up" account to me! Dial in, do what you want
> to do (get mail, upload my news posts, download other news posts, do
> whatever browsing) then get off-line to read your mail/news.

You start to get into the minutia of how things work.

Normal dial up usually ends up with a type of connection that is not
conducive to push things to; email, news, files, etc. So most dial up
is inherently pull methodologies wherein you send a request and the
reply comes back.

Conversely, UUCP, and as I understand some FTN methods, things that
would be pushed to you are queued at the system that you are calling
into waiting for you to connect and retrieve them.

In some ways the UUCP / QWK (?) methodologies are very much like curb
side pickup at stores. You pull up, notify the store that you are
there, and they bring everything out to you. Conversely traditional
dial up (and general Internet access) is more akin to you parking, going
in to find everything yourself, checking out before going back to your
car. The amount of time spent at / in the store is considerably
different even if you can end up with quite similar selection of things.

[Sn!pe]

Your post should have included modem sounds, "boing-boing" !

--
^Ï^. Sn!pe, PA, FIBS - Professional Crastinator

My pet rock Gordon just is.

[Paul Edwards]

On 01/01/24 01:17, Grant Taylor wrote:

> On 12/31/23 07:50, Paul Edwards wrote:
>> The trouble is if you haven't done it legally then you won't be able
>> to market it in the west, so the development effort is wasted.
>
> Either I'm not understanding you -or- you keep moving the goal post.
>
> The companies that have the micro-patches that I'm talking about are in
> business selling them for profit.

Micropatches are not remotely a competitor to Windows.
That is my interest.

>> I said "portable". ANSI X3.64 is the standard, not 3270. So I'm
>> looking to find out why people aren't following the standard.
>
> The wonderful thing about standards is that we have so many to choose from.
>
> 3270 /is/ a standard. Just not the standard that you want to use.

Sure - that would be the other approach - get
all PCs to do 3270 emulation instead.

PCs are indeed capable of being 3270 emulators,
and no-one disputes that.

But that prevents character-at-a-time input,
which is popular.

People disputed that character-at-a-time input
was possible on an IBM mainframe. That it was
not possible to do an EBCDIC version of
ANSI X3.64. And indeed, there was nothing to
point to to prove otherwise. Now there is.

>> DOS/VS is just a name. There's no point mentioning it just because the
>> names are coincidental. They don't even own that anymore anyway.
>
> I question the veracity that last statement.

DOS/VS (latest called z/VSE) was sold to
21st Century Software and is now called VSEn.

>> Their main product has always been MVS (now z/OS).
>
> For something to be the /main/ product implies there is at least one
> other product.

Sure. z/VM is their other mainframe offering.
There is z/TPF too, but that's very obscure/
limited use.

BFN. Paul.

[Paul Edwards]

On 01/01/24 18:14, Eric Pozharski wrote:
> with <umrriq$1oalv$1...@dont-email.me> Paul Edwards wrote:
>> On 31/12/23 09:52, Grant Taylor wrote:
>
>>>> Ok. And even if Americans can make something, if the Philippines is
>>>> under sanctions for some reason, can the Philippines independently
>>>> make it?
>>>
>>> Capability and technical acumen are completely independent of
>>> legality. Russia has had the former for many decades and did make /
>>> clone things without the latter.
>>
>> The trouble is if you haven't done it legally then you won't be able
>> to market it in the west, so the development effort is wasted.
>>
>> That's my interest in these rogue nations. I expect them to eventually
>> become non-rogue, and then I want that development effort to be
>> sellable.
>
> Excuse me, what is it you're smoking? I'm a bit scared, what if it's
> contageous?
>
> Rogue nations don't work this way. Like at all.

Nonsense. Both China and North Korea have
derivatives of Linux.

> Your ideas of how rogue nations function are probably taken out of ass.

Nope. Just an observation that both of them chose
to pick up a product that they will never be able
to legally close-source if that is deemed necessary
to produce a commercial competitor to Windows.

If they had written an OS from scratch, or picked
up something that was public domain, then they
would have complete control of their product, and
any subsequent development effort could at least
potentially compete with Microsoft.

They can in fact already compete with Microsoft
with their current product, but they will forever
be hamstrung by the fact that if they need to
close-source their product to protect their work -
which almost all software vendors decide is
appropriate for success (their sovereign decision -
even if you think they're wrong) they will be stuck.

Not sure what relevance the quote from Lenin was
about. Lenin wasn't involved in the software industry
from what I can remember. Xi and Kim both preside
over exactly that though. Both dead ends in my
opinion (which I acknowledge may be different from
your opinion, which is why I have no expectations
of you getting involved with PDOS).

BFN. Paul.

[Paul Edwards]

On 29/12/23 19:27, Paul Edwards wrote:

>> There are UUCP packages for Windows, MSDOS and OS/2 available at
>>
>> https://www.uupc.net/pub/uupc/prod/
>>
>> but these seem to be rather old (2002). The source code is also
>> available on the site.
>
> I do my development on Windows 2000 (last version
> that doesn't need "verification" or whatever they
> call it

"activation" I think it is.

> D:\devel\uupc>nmake -f nmake.mak
>
> and after a lot of weird warnings (it considers
> strcpy obsolete?), it completed successfully.
>
> uucico crashes deliberately (I checked the
> source code line reported) if there is any
> configuration error, so I did this:
>
> set UUPCSYSRC=fred.cfg
>
> and created a dummy config file, but it needs
> more than just a dummy.
>
> So assuming I am given an account I'll try to
> find out what it wants.

About 24 hours ago I was given a UUCP account (thanks!).

So now I get to start a new chapter of my life.

Here is what I see:

kerravon@kerravon-pc:~$ telnet uucp.eternal-september.org 540
Trying 95.217.65.141...
Connected to uucp.eternal-september.org.
Escape character is '^]'.
login: xxxx
Password:yyyyy
Shere=uucp.eternal-september.org
^]
telnet> quit
Connection closed.
kerravon@kerravon-pc:~$

I have suppressed my login and password
(both of which were provided to me - I
didn't choose them).

The login is different from the name I
use to connect to the nntp news server.
It's not very secret - I think it will
be obvious when I start posting. But I
suppressed it anyway.

So at last I see signs of life. The "Shere".
And I have a wonderful protocol to code
towards.

Not that wonderful actually. Fidonet (Wazoo)
is better. It starts off in binary and is
expecting to have to get past line noise.

This one looks like it won't cope with line
noise, and a will have a dorky script to log
in. You probably can retry the login even in
the face of line noise, but probably limited
attempts and not protected by checksum/CRC.

But the important thing is that I can see a
protocol and understand it and search for it
both online and in the source code I now have.

In order to implement this protocol with my
own independent (and presumably public domain)
software, I would use qemu to point COM1 to
this address and then hammer away at the serial
port, opened with fopen r+b and doing an
fseek 0 from SEEK_CUR when switching between
read and write. ie I can write standard C90
code. "com1:" will be a parameter so it is a
normal file operation.

ie everything is how I want it. And I already
have that working for nntp. ie I do my nntp
over com1: (before I started using the much
friendlier Thunderbird).

Now I know I need to clear my pending UUCP
batches before my replacement software is
ready (which could take months/years - or
maybe I will decide not to replace it at all,
and just write the UUCP to Fidonet software
I need, as it is the Fidonet software I wish
to follow in the long term for offline use,
and if the internet disappears for some reason,
I will no longer have a reason or the ability
to use UUCP). As my bundles are taking up space
on someone else's system and they won't be
happy if I don't regularly clear them (that's
exactly the position I want to be in).

So I will now begin the process of setting up
UUPC for the short/medium/long term - now that
I have a login.

It is unclear to me at this stage whether I need
to set it up to go over the COM port, or whether
I can use TCP/IP. I saw in the UUPC software that
there is a winsock directory, but so far the only
documentation I have read involved modems.

So I will report what I find out/progress.

BFN. Paul.

[immibis]

On 12/28/23 13:46, Paul Edwards wrote:
> 3. Just to clarify - this is for ordinary
> end users who wish to do UUCP, right? You
> don't have to be one of these always-on
> "peers", right?

I wasn't there when the old magic was written, as some were, but I
understand that UUCP is for batch transfers, and doesn't require
always-on access as you might expect over the Internet. Just like
Fidonet, the system was designed for machines (servers) to dial each
other every night, and do a bulk transfer in both directions.

I also understand that Fidonet points were only a necessity because of
Fidonet's authoritarian addressing scheme. And UUCP was designed by
people who had short serial links permanently connecting their terminals
to big timesharing systems, not people who used dial-up modems to
connect their microcomputers to BBS servers, so there was no "last mile"
problem.

You previously mentioned "bulk, efficient message transfer". A UUCP
peering doesn't offer that any more than an NNTP peering does, or even
just an NNTP account. The whole usenet is a store-and-forward network
and you can hold articles on your computer before posting them to any
server that lets you. You can also have your own server if you want, and
it doesn't have to be permanently connected (but why wouldn't you want
it to be?).

A UUCP exchange probably involves less network round trips than NNTP,
but I think NNTP has pipelining which would alleviate that.

UUCP is a historical curiosity, still operated as a hobby, not a
practical way to get news. Even NNTP is pretty antiquated, but at least
it's designed for the internet and not a dial-up modem working over the
analog phone system.

Usenet is a network of point-to-point links, transferring files we call
articles or messages. You can write an article on your computer and not
send it immediately, or you can send it immediately by NNTP, or you can
put it in a UUCP outbox, or whatever you want to do with it. You don't
need any special permission or protocol for any of that.

You can even invent your own protocol for the link you want to add, but
you'd better convince the other end of the link to install or write the
software to speak it to you.



If you're really keen on this disaster resilience stuff you wrote about
in another article, I'd be looking into amateur radio, maybe even
setting up some kind of usenet to radio gateway system, whether that's a
daily bulk broadcast or something interactive. A radio modem is a lot
like a telephone modem, so you might even find it convenient to use UUCP.
Note that when using radio you need a licence and you need to comply
with the applicable regulations, including a general ban on encrypted
and/or commercial traffic on amateur frequencies. I don't know the rest.
Also note that a lot of amateur radio boomers are also usenet boomers,
and there seem to be some groups with traffic under rec.radio.amateur.

I also notice an article <s5i9bu$oip$1...@gioia.aioe.org> from Chris Baird
in aus.radio.amateur.digital where he says he wants to set up some kind
of amateur radio netnews system he calls QRPBBB. However, it's from 2021.

[Paul Edwards]

On 06/01/24 17:26, Paul Edwards wrote:

> So I will report what I find out/progress.

Ok, so I have downloaded these already:

-rw-rw-r-- 1 kerravon kerravon 269970 Dec 29 18:54 upc13ks1.zip
-rw-rw-r-- 1 kerravon kerravon 248874 Dec 29 18:54 upc13ks2.zip
-rw-rw-r-- 1 kerravon kerravon 239145 Dec 29 18:54 upc13ks3.zip
-rw-rw-r-- 1 kerravon kerravon 180384 Dec 29 18:54 upc13ks4.zip
-rw-rw-r-- 1 kerravon kerravon 248543 Dec 29 21:00 upc13kad.zip

and since I am likely to be the person who
supports this product, I am importing the
source code into CVS:

D:\scratch\zzz>md uupc

D:\scratch\zzz>cd uupc

D:\scratch\zzz\uupc>unzip -n ..\upc13ks1.zip
D:\scratch\zzz\uupc>unzip -n ..\upc13ks2.zip
D:\scratch\zzz\uupc>unzip -n ..\upc13ks3.zip
D:\scratch\zzz\uupc>unzip -n ..\upc13ks4.zip

D:\scratch\zzz\uupc>cvs import -k o -m "release 1.13k import" uupc
kendra release-1_13k

Check out source code:

D:\devel>cvs get uupc

Open a visual studio 2005 professional command prompt (32-bit) then

D:\devel\uupc>nmake -f nmake.mak

Took about 2 minutes to build. Quite a lot
of seemingly spurious warnings, like this:

d:\devel\uupc\mail\deliver.c(1044) : warning C4996: 'strcpy': This
function or variable may be unsafe. Consider using strcpy_s instead. To
disable deprecation,
use _CRT_SECURE_NO_WARNINGS. See online help for details.


Seriously? You're going to warn about strcpy?


End result is:

D:\devel\uupc>dir *.exe *.dll
Volume in drive D is PAULDATA
Volume Serial Number is DAB2-6E30

Directory of D:\devel\uupc

2024-01-07 06:47a 200,704 expire.exe
2024-01-07 06:47a 135,168 fmt.exe
2024-01-07 06:47a 167,936 fromwho.exe
2024-01-07 06:47a 217,088 genhist.exe
2024-01-07 06:47a 167,936 gensig.exe
2024-01-07 06:47a 188,416 inews.exe
2024-01-07 06:47a 258,048 mail.exe
2024-01-07 06:47a 241,664 newsrun.exe
2024-01-07 06:47a 131,072 novrstrk.exe
2024-01-07 06:47a 172,032 regsetup.exe
2024-01-07 06:47a 258,048 rmail.exe
2024-01-07 06:47a 184,320 rnews.exe
2024-01-07 06:47a 204,800 sendbats.exe
2024-01-07 06:47a 319,488 uucico.exe
2024-01-07 06:47a 229,376 uucp.exe
2024-01-07 06:47a 176,128 uuname.exe
2024-01-07 06:47a 184,320 uupoll.exe
2024-01-07 06:48a 237,568 uupopd.exe
2024-01-07 06:47a 167,936 uuport.exe
2024-01-07 06:48a 258,048 uusmtpd.exe
2024-01-07 06:47a 229,376 uustat.exe
2024-01-07 06:47a 180,224 uusub.exe
2024-01-07 06:47a 237,568 uutraf.exe
2024-01-07 06:47a 204,800 uux.exe
2024-01-07 06:47a 245,760 uuxqt.exe

Directory of D:\devel\uupc

2024-01-07 06:47a 122,880 uupcdll.dll
26 File(s) 5,320,704 bytes
0 Dir(s) 7,903,203,328 bytes free

D:\devel\uupc>


While the size of those executables may look
extremely small for people today who download
a 100 MB package to do the most trivial of
tasks, they looked larger than I expected.

I took a look at one of the commands:

cl -c -DCRTAPI1=_cdecl -DCRTAPI2=_cdecl -nologo -GS -D_X86_=1
-DWIN32 -
D_WIN32 -W4 -D_WINNT -D_WIN32_WINNT=0x0500 -D_WIN32_IE=0x0500
-DWINVER=0x0500 -
D_MT -MTd -nologo -MT -Zi -Od -DDEBUG -DUDEBUG -Ilib
-FoC:\DOCUME~1\PAULED~1\L
OCALS~1\Tempobjnc\memstr.obj mail\memstr.c
cl : Command line warning D9025 : overriding '/MTd' with '/MT'
memstr.c

and the -DDEBUG suggests that it is doing a
dev build instead of release build by default.
-Od shows that too. Instead of -O2 or -Ox or
something. But that would have increased
compile time as well.

I had a brief look at nmake.mak but it didn't
obviously document which of these I should use
for production use, so I just moved on with my life:

nmk /nologo /s /f nmake.mak prodr
nmake /nologo /s /f nmake.mak prodp
nmake /nologo /s /f nmake.mak prodn


Hang on ... I think I can guess - r = real
mode, n = 16-bit NE executable, p = protected
mode or PE. Let me just try that.

I'll delete my dev source tree and check out
from CVS again to get a clean environment.

D:\devel\uupc>nmake /nologo /s /f nmake.mak prodp
NMAKE : fatal error U1073: don't know how to make 'prodp'
Stop.

D:\devel\uupc>

Um, yeah, thanks but no thanks.

Luckily I renamed the old directory. I'll just
go back to that, as my priority is to get this
to work at all. If I were distributing this I
would organize optimization and static linking
probably, but so far I haven't made any code
changes.

Ok, so I now need to find some documentation
because I can't guess the config files.

ie I know already I need to do something like:

set UUPCSYSRC=fred.cfg

And I presumably downloaded this:

-rw-rw-r-- 1 kerravon kerravon 248543 Dec 29 21:00 upc13kad.zip

for a reason. And that reason is?

upc13kad.zip 2002-01-07 17:35 243K Documentation and Samples (Plain text
format)


Sounds good to me. Plain text? Command line? That's me!

>>> There are UUCP packages for Windows, MSDOS and OS/2 available at
>>>
>>> https://www.uupc.net/pub/uupc/prod/

Manual is here:

https://www.uupc.net/UUPC-Manual.htm

(I do read documentation online rather
than using text files). I might try
seeing if there is an appropriate sample
file rather than reading documentation
though.

It will take a while to figure this out,
so as Arnie said - "I'll be back".

BFN. Paul.

[Paul Edwards]

On 07/01/24 07:20, Paul Edwards wrote:

> and since I am likely to be the person who
> supports this product, I am importing the
> source code into CVS:

Ok, so I have my first (likely) bug.

I put in my proper userid and password and
it connected, but I think it didn't like
the Shere= from eternal-september and showed
garbage. I put in a fake userid and password,
and I still got the garbage:

D:\uupc>doit

D:\uupc>set UUPCSYSRC=uupc.rc

D:\uupc>set UUPCUSRRC=kerravon.rc

D:\uupc>rem even though philippines is 8 hours ahead, use negative 8 instead

D:\uupc>set TZ=PST-8

D:\uupc>echo machine=uucp.eternal-september.org 1>permissn

D:\uupc>uucico -s uucp.eternal-september.org
uucico: UUPC/extended 1.13k (Win32 mode, 7Jan24 06:47) [debug enabled]
Changes and Compilation Copyright (c) 1989-2002 by Kendra Electronic
Wonderworks. May be freely distributed for reasonable copying fee
if original documentation and source is included. See license for
details and restrictions.
callup: Connecting to uucp.eternal-septembÉ @ via TCPIP on Sun Jan 07
08:42:19 2024
Connecting to host uucp.eternal-september.org [95.217.65.141] port 540
tsread: EOF on recv()
Startup: Timeout for first message
tsread: EOF on recv()
Could not connect to remote system.
copylog: Spun off log D:\uupc\log\uupc0409.log to D:/uupc/log/uucico.log

D:\uupc>


It's probably a restriction on the length of the
host because I've probably not defined things
properly, because so far I have read minimal
documentation and done maximum guessing.

So let me look at the source code first to try
to find out where they have presumably failed
to check for buffers being exceeded.

BFN. Paul.




D:\uupc>type doit.bat
set UUPCSYSRC=uupc.rc
set UUPCUSRRC=kerravon.rc

rem even though philippines is 8 hours ahead, use negative 8 instead
set TZ=PST-8

echo machine=uucp.eternal-september.org > permissn

uucico -s uucp.eternal-september.org

D:\uupc>type uupc.rc
# *--------------------------------------------------------------------*
# * Sample system configuration for toscis. See *
# * UUPC/extended installation documentation for use. *
# *--------------------------------------------------------------------*
Version=1.13k
NodeName=xxxx
Domain=xxxx.eternal-september.org
postmaster=xxxx
Mailserv=xxxx
#InModem=TB2500

# *--------------------------------------------------------------------*
# * The follow global defaults are suggested *
# *--------------------------------------------------------------------*
#DOS.Editor=Edit %s
#OS2.Editor=E %s
#Win16.Editor=notepad %s
#Win32.Editor=notepad %s

#MailExt=SPB

D:\uupc>type systems
# Default port is 540. Not sure if there is a way of
# changing that if the host requires it

uucp.eternal-september.org Any TCPIP 38400 uucp.eternal-september.org et
login: xxxx\ryyyy

D:\uupc>

[Paul Edwards]

On 07/01/24 07:08, immibis wrote:

> You previously mentioned "bulk, efficient message transfer". A UUCP
> peering doesn't offer that any more than an NNTP peering does, or even
> just an NNTP account.

I have no idea why you would say that,
which is inconsistent with where you say:

> UUCP is a historical curiosity, still operated as a hobby, not a
> practical way to get news. Even NNTP is pretty antiquated, but at least
> it's designed for the internet and not a dial-up modem working over the
> analog phone system.

Why do you think UUCP is better suited to NNTP
when using a dial-up modem?

That's the answer to your question.

And that's the exact thing I wish to optimize for.

Just because I happen to be on a permanently-connected
800 Mbps fiber-optic link *right now*, I wish to be
prepared for not just dialup, but sneakernet.

> If you're really keen on this disaster resilience stuff you wrote about
> in another article, I'd be looking into amateur radio, maybe even

It doesn't need to be me personally who does
a link that requires more than walking, or
riding on a horse.

My job ends when I have a zip file on a
floppy disk or USB stick, to attach to
the saddle of a horse or the leg of a pigeon,
or any other means of transport.

BFN. Paul.

[Jon Ribbens]

On 2024-01-07, Paul Edwards <muta...@gmail.com> wrote:
> On 07/01/24 07:08, immibis wrote:
>> You previously mentioned "bulk, efficient message transfer". A UUCP
>> peering doesn't offer that any more than an NNTP peering does, or even
>> just an NNTP account.
>
> I have no idea why you would say that,
> which is inconsistent with where you say:
>
>> UUCP is a historical curiosity, still operated as a hobby, not a
>> practical way to get news. Even NNTP is pretty antiquated, but at least
>> it's designed for the internet and not a dial-up modem working over the
>> analog phone system.
>
> Why do you think UUCP is better suited to NNTP
> when using a dial-up modem?
>
> That's the answer to your question.
>
> And that's the exact thing I wish to optimize for.

Some of us where alive in the 1990s and did exactly this (offline Usenet
systems via dialup) and NNTP copes fine, provided that the server supports
NEWNEWS efficiently.

[Paul Edwards]

On 07/01/24 10:40, Jon Ribbens wrote:

>> Why do you think UUCP is better suited to NNTP
>> when using a dial-up modem?
>>
>> That's the answer to your question.
>>
>> And that's the exact thing I wish to optimize for.
>
> Some of us where alive in the 1990s and did exactly this (offline Usenet
> systems via dialup) and NNTP copes fine, provided that the server supports
> NEWNEWS efficiently.

And what if both ends were Commodore 64s
in the 1980s using a 300 bps modem and you
were dialing internationally?

NEWNEWS presumably can't predict that you're
about to request the exact same newsgroups
that you got last time, and have a compressed
bundle waiting there for you.

If it could do that, it would be called UUCP.

BFN. Paul.

[Paul Edwards]

On 07/01/24 08:45, Paul Edwards wrote:

> So let me look at the source code first to try
> to find out where they have presumably failed
> to check for buffers being exceeded.

D:\devel\uupc>cvs diff -c
cvs diff: Diffing .
cvs diff: Diffing LIB
cvs diff: Diffing MAIL
cvs diff: Diffing NEWS
cvs diff: Diffing SCRIPTS
cvs diff: Diffing TEST
cvs diff: Diffing UTIL
cvs diff: Diffing UUCICO
Index: UUCICO/dcpsys.c
===================================================================
RCS file: \cvsroot/uupc/UUCICO/dcpsys.c,v
retrieving revision 1.1.1.1
diff -c -r1.1.1.1 dcpsys.c
*** UUCICO/dcpsys.c 6 Jan 2024 22:42:33 -0000 1.1.1.1
--- UUCICO/dcpsys.c 7 Jan 2024 02:44:47 -0000
***************
*** 390,395 ****
--- 390,402 ----
}

strcpy(protocols, flds[FLD_PROTO]);
+ if (strlen(flds[FLD_REMOTE]) > (sizeof rmtname - 1))
+ {
+ fprintf(stderr,
+ "remote name too big - %d\n",
+ strlen(flds[FLD_REMOTE]));
+ exit(EXIT_FAILURE);
+ }
strcpy(rmtname, flds[FLD_REMOTE]);

/*--------------------------------------------------------------------*/
cvs diff: Diffing UUCP
cvs diff: Diffing UUTRAF
cvs diff: Diffing WINSOCK
cvs diff: Diffing docs
cvs diff: Diffing icons.os2
cvs diff: Diffing icons.win
cvs diff: Diffing winword

D:\devel\uupc>

D:\uupc>uucico -s uucp.eternal-september.org
uucico: UUPC/extended 1.13k (Win32 mode, 7Jan24 06:47) [debug enabled]
Changes and Compilation Copyright (c) 1989-2002 by Kendra Electronic
Wonderworks. May be freely distributed for reasonable copying fee
if original documentation and source is included. See license for
details and restrictions.

remote name too big - 26
copylog: Spun off log D:\uupc\log\uupc034d.log to D:/uupc/log/uucico.log

D:\uupc>



D:\devel\uupc\UUCICO>grep rmtname *.h
dcp.h: extern char rmtname[20]; /* system we end up talking to */

D:\devel\uupc\UUCICO>



So now I know where I stand on that issue.

I'll see what I can do.

BFN. Paul.

[immibis]

On 1/7/24 03:02, Paul Edwards wrote:
>
> Just because I happen to be on a permanently-connected
> 800 Mbps fiber-optic link *right now*, I wish to be
> prepared for not just dialup, but sneakernet.

But UUCP isn't sneakernet, either. What you want for a sneakernet is for
the server to dump plain news articles to a directory (or tape :) ),
then you manually bring the directory to the other server and read them in.

You'll probably find UUCP spooling useful to hack such a system, because
sneakernet is batch communication just like UUCP, but you won't actually
use UUCP, just the parts that surround it.

>
>> If you're really keen on this disaster resilience stuff you wrote about
>> in another article, I'd be looking into amateur radio, maybe even
>
> It doesn't need to be me personally who does
> a link that requires more than walking, or
> riding on a horse.
>
> My job ends when I have a zip file on a
> floppy disk or USB stick, to attach to
> the saddle of a horse or the leg of a pigeon,
> or any other means of transport.
>

Why set your limit there? On the abstraction spectrum, from sending
files, to sending USB drives, to sending actual data signals, why stop
at USB drives?

[immibis]

On 1/7/24 03:49, Paul Edwards wrote:
> On 07/01/24 10:40, Jon Ribbens wrote:
>
>>> Why do you think UUCP is better suited to NNTP
>>> when using a dial-up modem?
>>>
>>> That's the answer to your question.
>>>
>>> And that's the exact thing I wish to optimize for.
>>
>> Some of us where alive in the 1990s and did exactly this (offline Usenet
>> systems via dialup) and NNTP copes fine, provided that the server
>> supports
>> NEWNEWS efficiently.
>
> And what if both ends were Commodore 64s
> in the 1980s using a 300 bps modem and you
> were dialing internationally?
>

Then it's going to take a long time no matter what you do, and you're
bottlenecked on the actual data transmission anyway.

Why would you use unix-to-unix-copy-protocol for c64-to-c64 copying?
You'd use something optimized for c64-to-c64. Probably using a
rudimentary compression algorithm, extremely sprase overhead, and no
base64 encoding.

This part of the Usenet technology stack is hop-by-hop, so each link can
use a technology that best suits it.

> NEWNEWS presumably can't predict that you're
> about to request the exact same newsgroups
> that you got last time, and have a compressed
> bundle waiting there for you.

If you only have a 300 bps modem, you *really* don't want to transmit
messages the recipient already has. NEWNEWS works like this:
The client tells the server which groups it's interested in and the last
time it received updates.
The server tells the client a list of new messages in these groups since
this time.
The client tells the server which messages it wants to receive.
The server sends the messages to the client.

Note that these commands can be pipelined, so only a few round trip
times are needed, regardless of the number of new messages.

There's an NNTP streaming mode (RFC 4644) which allows a server to just
send new articles (typically to another server) without any
back-and-forth. This is the mode you'd want for your binary groups
peering on your 800mbps fiber link, or a sneakernet link (if you could
use NNTP), or IPoAC. It is a stupid waste of bandwidth for a 300bps
modem link.

Also, get a better modem.

[Jon Ribbens]

On 2024-01-07, Paul Edwards <muta...@gmail.com> wrote:

> On 07/01/24 10:40, Jon Ribbens wrote:
>> Some of us where alive in the 1990s and did exactly this (offline Usenet
>> systems via dialup) and NNTP copes fine, provided that the server supports
>> NEWNEWS efficiently.
>
> And what if both ends were Commodore 64s
> in the 1980s using a 300 bps modem and you
> were dialing internationally?

300 bps? Commodore 64? Luxury.
Lower your voice, your talking is interfering with my acoustic coupler.
Young people today don't know they're born etc.

> NEWNEWS presumably can't predict that you're
> about to request the exact same newsgroups
> that you got last time, and have a compressed
> bundle waiting there for you.

Why do you think I said "efficiently"?

If you're worried about compression, did we un-invent PPP CCP and
NNTP COMPRESS while I wasn't looking?

[Paul Edwards]

On 07/01/24 10:54, Paul Edwards wrote:

> I'll see what I can do.

D:\uupc>doit

D:\uupc>set UUPCSYSRC=uupc.rc

D:\uupc>set UUPCUSRRC=kerravon.rc

D:\uupc>rem even though philippines is 8 hours ahead, use negative 8 instead

D:\uupc>set TZ=PST-8

D:\uupc>echo machine=eternal 1>permissn

D:\uupc>rem uucico -s uucp.eternal-september.org

D:\uupc>uucico -s eternal

uucico: UUPC/extended 1.13k (Win32 mode, 7Jan24 06:47) [debug enabled]
Changes and Compilation Copyright (c) 1989-2002 by Kendra Electronic
Wonderworks. May be freely distributed for reasonable copying fee
if original documentation and source is included. See license for
details and restrictions.

callup: Connecting to eternal via TCPIP on Sun Jan 07 10:59:26 2024

Connecting to host uucp.eternal-september.org [95.217.65.141] port 540
tsread: EOF on recv()
Startup: Timeout for first message
tsread: EOF on recv()

0 files sent, 0 files received, 0 bytes sent, 0 bytes received
0 packets transferred, 0 errors, connection time 0:02, 0 bytes/second

Could not connect to remote system.

copylog: Spun off log D:\uupc\log\uupc0111.log to D:/uupc/log/uucico.log

D:\uupc>type systems
# Default port is 540. Not sure if there is a way of
# changing that if the host requires it

eternal Any TCPIP 38400 uucp.eternal-september.org et login: xxxx\ryyyy

D:\uupc>



Ok, so putting in my proper userid and password, I get:

Startup: Wrong host uucp.eternal-september.org, expected eternal

which is what had prompted me to put in the full name
in the first place.


I'm not sure what the right thing to do about that is.
So for now I'll try just changing the source code:

D:\devel\uupc\UUCICO>cvs diff dcp.c dcp.h
Index: dcp.c
===================================================================
RCS file: \cvsroot/uupc/UUCICO/dcp.c,v
retrieving revision 1.1.1.1
diff -r1.1.1.1 dcp.c
308c308

< char rmtname[20]; /* system we end up talking to */

---
> char rmtname[30]; /* system we end up talking to */
Index: dcp.h
===================================================================
RCS file: \cvsroot/uupc/UUCICO/dcp.h,v
retrieving revision 1.1.1.1
diff -r1.1.1.1 dcp.h
172c172

< extern char rmtname[20]; /* system we end up talking to */

---
> extern char rmtname[30]; /* system we end up talking to */

D:\devel\uupc\UUCICO>



And the result is ...

D:\uupc>doit

D:\uupc>set UUPCSYSRC=uupc.rc

D:\uupc>set UUPCUSRRC=kerravon.rc

D:\uupc>rem even though philippines is 8 hours ahead, use negative 8 instead

D:\uupc>set TZ=PST-8

D:\uupc>echo machine=uucp.eternal-september.org 1>permissn

D:\uupc>uucico -s uucp.eternal-september.org
uucico: UUPC/extended 1.13k (Win32 mode, 7Jan24 06:47) [debug enabled]
Changes and Compilation Copyright (c) 1989-2002 by Kendra Electronic
Wonderworks. May be freely distributed for reasonable copying fee
if original documentation and source is included. See license for
details and restrictions.

callup: Connecting to uucp.eternal-september.org via TCPIP on Sun Jan 07
11:07:38 2024

Connecting to host uucp.eternal-september.org [95.217.65.141] port 540

Unexpected second message: You are unknown to me
tsread: EOF on recv()
0 files sent, 0 files received, 0 bytes sent, 0 bytes received
0 packets transferred, 0 errors, connection time 0:03, 0 bytes/second

Could not connect to remote system.

copylog: Spun off log D:\uupc\log\uupc02b1.log to D:/uupc/log/uucico.log

D:\uupc>


So I'll try changing some (specifically the NodeName
and Domain) of these xxxx to my system name
(given to me by eternal september):

D:\uupc>type uupc.rc
# *--------------------------------------------------------------------*
# * Sample system configuration for toscis. See *
# * UUPC/extended installation documentation for use. *
# *--------------------------------------------------------------------*
Version=1.13k
NodeName=xxxx
Domain=xxxx.eternal-september.org
postmaster=xxxx
Mailserv=xxxx
#InModem=TB2500

# *--------------------------------------------------------------------*
# * The follow global defaults are suggested *
# *--------------------------------------------------------------------*
#DOS.Editor=Edit %s
#OS2.Editor=E %s
#Win16.Editor=notepad %s
#Win32.Editor=notepad %s

#MailExt=SPB

D:\uupc>

and ... hallelujah!

D:\uupc>uucico -s uucp.eternal-september.org
uucico: UUPC/extended 1.13k (Win32 mode, 7Jan24 06:47) [debug enabled]
Changes and Compilation Copyright (c) 1989-2002 by Kendra Electronic
Wonderworks. May be freely distributed for reasonable copying fee
if original documentation and source is included. See license for
details and restrictions.

callup: Connecting to uucp.eternal-september.org via TCPIP on Sun Jan 07
11:10:19 2024

Connecting to host uucp.eternal-september.org [95.217.65.141] port 540

(CENSORED-MY-HOSTNAME) connected to uucp.eternal-september.org: network
link, e protocol, z grade
Receiving "D.001I" as "D.uucp.etd001I" (uucp.ete/D/6)gh0jp)
Receiving "D.X001T" as "X.uucp.etN001T" (uucp.ete/X/6&'))t0)
Receiving "D.001D" as "D.uucp.etd001D" (uucp.ete/D/6)gh0jk)
Receiving "D.X001U" as "X.uucp.etN001U" (uucp.ete/X/6&'))t1)
Receiving "D.001E" as "D.uucp.etd001E" (uucp.ete/D/6)gh0jl)
Receiving "D.X001V" as "X.uucp.etN001V" (uucp.ete/X/6&'))t2)
Receiving "D.001F" as "D.uucp.etd001F" (uucp.ete/D/6)gh0jm)
Receiving "D.X001W" as "X.uucp.etN001W" (uucp.ete/X/6&'))t3)
Receiving "D.001G" as "D.uucp.etd001G" (uucp.ete/D/6)gh0jn)
Receiving "D.X001X" as "X.uucp.etN001X" (uucp.ete/X/6&'))t4)
Receiving "D.000S" as "D.uucp.etd000S" (uucp.ete/D/6)gh0i3)
Receiving "D.X001Y" as "X.uucp.etN001Y" (uucp.ete/X/6&'))t5)
Receiving "D.001H" as "D.uucp.etd001H" (uucp.ete/D/6)gh0jo)
Receiving "D.X001Z" as "X.uucp.etN001Z" (uucp.ete/X/6&'))t6)
Receiving "D.000V" as "D.uucp.etd000V" (uucp.ete/D/6)gh0i6)
Receiving "D.X0020" as "X.uucp.etN0020" (uucp.ete/X/6&'))u0)
Receiving "D.000T" as "D.uucp.etd000T" (uucp.ete/D/6)gh0i4)
Receiving "D.X0021" as "X.uucp.etN0021" (uucp.ete/X/6&'))u1)
Receiving "D.001J" as "D.uucp.etd001J" (uucp.ete/D/6)gh0jq)
Receiving "D.X0022" as "X.uucp.etN0022" (uucp.ete/X/6&'))u2)
Receiving "D.001K" as "D.uucp.etd001K" (uucp.ete/D/6)gh0jr)
Receiving "D.X0023" as "X.uucp.etN0023" (uucp.ete/X/6&'))u3)
Receiving "D.001L" as "D.uucp.etd001L" (uucp.ete/D/6)gh0js)
Receiving "D.X0024" as "X.uucp.etN0024" (uucp.ete/X/6&'))u4)

...

Receiving "D.X003G" as "X.uucp.etN003G" (uucp.ete/X/6&'))w7)
Receiving "D.001A" as "D.uucp.etd001A" (uucp.ete/D/6)gh0jh)
Receiving "D.X003H" as "X.uucp.etN003H" (uucp.ete/X/6&'))w8)
Receiving "D.000R" as "D.uucp.etd000R" (uucp.ete/D/6)gh0i2)
Receiving "D.X003I" as "X.uucp.etN003I" (uucp.ete/X/6&'))w9)
Receiving "D.001B" as "D.uucp.etd001B" (uucp.ete/D/6)gh0ji)
Receiving "D.X003J" as "X.uucp.etN003J" (uucp.ete/X/6&'))w@)
Receiving "D.001C" as "D.uucp.etd001C" (uucp.ete/D/6)gh0jj)
Receiving "D.X003K" as "X.uucp.etN003K" (uucp.ete/X/6&'))w^)
0 files sent, 128 files received, 773 bytes sent, 308843 bytes received
904 packets transferred, 0 errors, connection time 2:03, 2517 bytes/second
copylog: Spun off log D:\uupc\log\uupc0519.log to D:/uupc/log/uucico.log

D:\uupc>


That's quite a lot of files for a couple of days.

And these are very unfriendly filenames:

D:\uupc\spool>dir /s /p

Volume in drive D is PAULDATA
Volume Serial Number is DAB2-6E30

Directory of D:\uupc\spool

2024-01-07 11:12a <DIR> .
2024-01-07 11:12a <DIR> ..
2024-01-07 11:12a <DIR> locks.lck
2024-01-07 11:10a <DIR> uucp.ete
0 File(s) 0 bytes

Directory of D:\uupc\spool\locks.lck

2024-01-07 11:12a <DIR> .
2024-01-07 11:12a <DIR> ..
0 File(s) 0 bytes

Directory of D:\uupc\spool\uucp.ete

2024-01-07 11:10a <DIR> .
2024-01-07 11:10a <DIR> ..
2024-01-07 11:12a <DIR> D
2024-01-07 11:12a <DIR> X
0 File(s) 0 bytes

Directory of D:\uupc\spool\uucp.ete\D

2024-01-07 11:12a <DIR> .
2024-01-07 11:12a <DIR> ..
2024-01-07 11:11a 118,417 6)gh0hh
2024-01-07 11:12a 1,360 6)gh0hi
2024-01-07 11:11a 6,208 6)gh0hj
2024-01-07 11:11a 1,841 6)gh0hk
2024-01-07 11:11a 1,942 6)gh0hl
2024-01-07 11:11a 1,300 6)gh0hm
2024-01-07 11:11a 4,100 6)gh0hn
2024-01-07 11:11a 10,892 6)gh0ho
...


What will they do next? Put spaces in filenames?!

...

Directory of D:\uupc\spool\uucp.ete\X

2024-01-07 11:12a <DIR> .
2024-01-07 11:12a <DIR> ..
2024-01-07 11:10a 78 6&'))t0
2024-01-07 11:10a 78 6&'))t1
2024-01-07 11:10a 78 6&'))t2
2024-01-07 11:10a 78 6&'))t3
2024-01-07 11:10a 78 6&'))t4
...
2024-01-07 11:12a 78 6&'))w7
2024-01-07 11:12a 78 6&'))w8
2024-01-07 11:12a 78 6&'))w9
2024-01-07 11:12a 78 6&'))w@
2024-01-07 11:12a 78 6&'))w^
64 File(s) 4,992 bytes

Total Files Listed:
128 File(s) 302,947 bytes


Now - does it play Doom?

2024-01-07 11:12a 1,360 6)gh0hi
2024-01-07 11:11a 1,300 6)gh0hm
2024-01-07 11:11a 1,261 6)gh0hz
2024-01-07 11:11a 1,212 6)gh0hx
2024-01-07 11:12a <DIR> ..
2024-01-07 11:12a <DIR> .
64 File(s) 297,955 bytes
2 Dir(s) 7,895,879,680 bytes free

D:\uupc\spool\uucp.ete\D>dir /o-s


D:\uupc\spool\uucp.ete\D>hexdump 6)gh0hx 0 200
000000 23212063 756E6261 7463680A 1F9D9023 #! cunbatch....#
000010 428090E3 A6CC9D39 2062D0B0 91430194 B......9 b...C..
000020 3074D0E8 0051864E 198261D8 B4985306 0t...Q.N..a...S.
000030 8EC53662 2EBA7823 E74C8882 075D54BC ..6b..x#.L...]T.
000040 E826E3C6 8E1F43CA 19593284 0B284FA6 .&....C..Y2..(O.
000050 502942E4 E41B3A2D CC906CD1 264C1A36 P)B...:-..l.&L.6
000060 0A8CC879 D366E29D 376CCC9C 09E3E60C ...y.f..7l......
000070 88352C41 F070F3E6 0E9A8B65 8014B423 .5,A.p.....e...#
000080 12A20F05 4E0CCE39 B3B40E9C 391333D2 ....N..9....9.3.
000090 1939C705 993264D9 BC81D3A6 8C1B3A0A .9...2d.......:.


ok, so I'm on Windows, not expecting to have
a cunbatch, nor the ability to run a shell script,
and 0D0A for line terminators is more normal too.

So - what are my options? 7z?

None of the *.exe files look like they will
decompress a single file for me.

And I'm not ready to run the full processing
either. I just want the raw data, so that when
the rest of my software suite is ready, I can
import months/years of data into it (given that
google groups will no longer have new data). I
intend to continue using Thunderbird in the
meantime.

But I just need some feedback that this compressed
data can be uncompressed, so that I know that I'm
not preserving defacto junk.

What is the compression algorithm being used, and
what are my options for decompressing it?

Let me see what I can do with that file I just
hexdumped.

BFN. Paul.

[Paul Edwards]

On 07/01/24 10:58, immibis wrote:
> On 1/7/24 03:02, Paul Edwards wrote:
>>
>> Just because I happen to be on a permanently-connected
>> 800 Mbps fiber-optic link *right now*, I wish to be
>> prepared for not just dialup, but sneakernet.
>
> But UUCP isn't sneakernet, either.

True. Note that I already have this working
with Fidonet technology. It's all on the
PDOS/386 distribution at http://pdos.org

And yes, at the last minute, instead of
using Wazoo over a dialup modem, you copy
the files that you had intended to send
over dialup modem, and copy them to floppy
disk or whatever instead.

So I wish to have the same setup with UUCP.
Yes, if you have a dialup available - great -
use it. If you have a 800 Mbps fiber-optic
link - even greater - use that instead. If
you don't, then sneakernet it.

> What you want for a sneakernet is for
> the server to dump plain news articles to a directory (or tape :) ),
> then you manually bring the directory to the other server and read them in.

Yes, especially if you zip it first. Absolutely
it can work.

But then you're not ready for dialup over an
international link, because the NNTP server
won't take your zip. Apparently the NNTP
server is capable of doing on-the-fly
compression. But who has time to wait online
on an international link for the 3 hours it
takes for a Commodore 64 to do that compression?

Or rather - who has the money?

> You'll probably find UUCP spooling useful to hack such a system, because
> sneakernet is batch communication just like UUCP, but you won't actually
> use UUCP, just the parts that surround it.

Sure.

And as I've said previously, the actual
sneakernet will be Fidonet anyway. I am
only going to use Usenet technology while
Usenet actually exists. And then use it
to immediately convert to Fidonet format
using UFGATE or my own software (I've never
researched options - I've never been here
before).

>>> If you're really keen on this disaster resilience stuff you wrote about
>>> in another article, I'd be looking into amateur radio, maybe even
>>
>> It doesn't need to be me personally who does
>> a link that requires more than walking, or
>> riding on a horse.
>>
>> My job ends when I have a zip file on a
>> floppy disk or USB stick, to attach to
>> the saddle of a horse or the leg of a pigeon,
>> or any other means of transport.
>
> Why set your limit there? On the abstraction spectrum, from sending
> files, to sending USB drives, to sending actual data signals, why stop
> at USB drives?

That's all a question of hardware.

I'm a software engineer. And I'm doing the bit
that applies no matter what hardware you are
using, rather than e.g. a driver for a HAM radio.

E.g. enhancements to msged to cope with the
Usenet headers. Who else is going to write that
if not me?

BFN. Paul.

[Paul Edwards]

On 07/01/24 11:06, immibis wrote:

>> And what if both ends were Commodore 64s
>> in the 1980s using a 300 bps modem and you
>> were dialing internationally?
>
> Then it's going to take a long time no matter what you do, and you're
> bottlenecked on the actual data transmission anyway.

Absolutely. That's the only thing that
matters at all - data transmission time.

> Why would you use unix-to-unix-copy-protocol for c64-to-c64 copying?
> You'd use something optimized for c64-to-c64. Probably using a
> rudimentary compression algorithm, extremely sprase overhead, and no
> base64 encoding.

I can't predict that the system I wish to connect
to has a C64-friendly protocol. UUCP is the standard.
And regardless - who is going to write this C64
protocol? Once again - UUCP actually exists.

Yes, in the future - when someone volunteers to
write the software (crickets all round), UUCP
can be replaced.

>> NEWNEWS presumably can't predict that you're
>> about to request the exact same newsgroups
>> that you got last time, and have a compressed
>> bundle waiting there for you.
>
> If you only have a 300 bps modem, you *really* don't want to transmit
> messages the recipient already has.

I expect my peer system to only provide
new news via UUCP. That's the whole purpose
of it. New news, that my system has never
received before, all compressed and waiting
for me. The peer knows exactly what I have
and what I want - it provided it all to me.

> Also, get a better modem.

I bought my first modem in 1987 I believe,
and it was 1200 bps and cost something like
$500.

Not everyone can afford that (or more
accurately - prioritize that).

I did indeed get a faster modem in the 1990s,
and one day my BBS modem died, and the only
working modem I had was that 1200 bps modem.

One of my users thought that something was
wrong and tried reconnecting multiple times
to get a faster speed, but eventually decided
to grin and bear it. I think that was an
expensive interstate (Australia) call.

Yes, he was desperate enough for text messages,
as we all were, to use a 1200 bps modem.

If I'd been born a few years earlier than 1967,
I would have been equally desperate over a
300 bps link.

BFN. Paul.

[Paul Edwards]

On 07/01/24 11:08, Jon Ribbens wrote:

>> NEWNEWS presumably can't predict that you're
>> about to request the exact same newsgroups
>> that you got last time, and have a compressed
>> bundle waiting there for you.
>
> Why do you think I said "efficiently"?
>
> If you're worried about compression, did we un-invent PPP CCP and
> NNTP COMPRESS while I wasn't looking?

You are forced to do that compression on the fly.

So what you are calling "efficient", I am
calling either "non-existent" or "3 hours
on an international call compared to UUCP".

I'm not saying that YOU have a need to do
any of this. If you have a fast server -
which - right now - you do - go right ahead
and use it.

I don't want to be dependent on that.

I want a solution for low-capability systems -
both ends.

One day I may be last man standing.

Or maybe I won't.

I don't particularly care. It's my hobby, and
this is a solution for the hobbyist Fidonet
network - even though I am no longer in it.
I used to be, and I'm still in the process
of completing the software suite I envisage.

BFN. Paul.

[immibis]

On 1/7/24 04:43, Paul Edwards wrote:
> On 07/01/24 10:58, immibis wrote:
>> On 1/7/24 03:02, Paul Edwards wrote:
>>>
>>> Just because I happen to be on a permanently-connected
>>> 800 Mbps fiber-optic link *right now*, I wish to be
>>> prepared for not just dialup, but sneakernet.
>>
>> But UUCP isn't sneakernet, either.
>
> True. Note that I already have this working
> with Fidonet technology. It's all on the
> PDOS/386 distribution at http://pdos.org
>

> And yes, at the last minute, instead of
> using Wazoo over a dialup modem, you copy
> the files that you had intended to send
> over dialup modem, and copy them to floppy
> disk or whatever instead.
>

Sure. It would be inaccurate to call the resulting system a "Wazoo link"
even though you've taken the workflow intended for Wazoo and made some
modifications - because the actual Wazoo part is stripped out.

Once upon a time, I solved a problem in an automated hardware testing
system by sending IP packets encoded in hexadecimal, to the equipment's
debug RS-232 port, through a TCP connection to a serial port
concentrator. It would be incorrect to call it an Ethernet link.

> So I wish to have the same setup with UUCP.
> Yes, if you have a dialup available - great -
> use it. If you have a 800 Mbps fiber-optic
> link - even greater - use that instead. If
> you don't, then sneakernet it.
>

Presumably the peers you can reach through 800Mbps fiber-optic Internet
aren't the same peers you can reach over sneakernet. You might be
solving this problem one layer too low.

(If the 800Mbps fiber-optic is just going over to the next building,
they might be.)

>> You'll probably find UUCP spooling useful to hack such a system, because
>> sneakernet is batch communication just like UUCP, but you won't actually
>> use UUCP, just the parts that surround it.
>
> Sure.
>
> And as I've said previously, the actual
> sneakernet will be Fidonet anyway.

then I don't understand the discussion about UUCP over sneakernet.

>
> That's all a question of hardware.
>
> I'm a software engineer. And I'm doing the bit
> that applies no matter what hardware you are
> using, rather than e.g. a driver for a HAM radio.

P.S. "ham" isn't an acronym. It apocryphally stems from new Morse code
operators having a "ham fist".

P.P.S. get an RTL-SDR and play with it a little.

P.P.P.S. transmitting data over sound card is a software thing :)

[immibis]

On 1/7/24 04:52, Paul Edwards wrote:
> On 07/01/24 11:06, immibis wrote:
>
> I can't predict that the system I wish to connect
> to has a C64-friendly protocol. UUCP is the standard.
> And regardless - who is going to write this C64
> protocol? Once again - UUCP actually exists.
>

NNTP is the standard, actually.

>>> NEWNEWS presumably can't predict that you're
>>> about to request the exact same newsgroups
>>> that you got last time, and have a compressed
>>> bundle waiting there for you.
>>
>> If you only have a 300 bps modem, you *really* don't want to transmit
>> messages the recipient already has.
>
> I expect my peer system to only provide
> new news via UUCP. That's the whole purpose
> of it. New news, that my system has never
> received before, all compressed and waiting
> for me. The peer knows exactly what I have
> and what I want - it provided it all to me.

Do you expect to have more than one peer?

[Ray Banana]

Thus spake Paul Edwards <muta...@gmail.com>

[...]

> ok, so I'm on Windows, not expecting to have
> a cunbatch, nor the ability to run a shell script,
> and 0D0A for line terminators is more normal too.

Remember that both U's mean UNIX, so 0A is the standard line end.

> So - what are my options? 7z?

In the source code you will find a file named rnews.c

All batches should be processed by rnews, it knows how to cunbatch
these input files.

HTH

--
Пу́тін — хуйло́
https://www.eternal-september.org

[SugarBug]

On Sun, 7 Jan 2024 11:52:38 +0800
Paul Edwards <muta...@gmail.com> wrote:

> I can't predict that the system I wish to connect
> to has a C64-friendly protocol. UUCP is the standard.
> And regardless - who is going to write this C64
> protocol? Once again - UUCP actually exists.
>
> Yes, in the future - when someone volunteers to
> write the software (crickets all round), UUCP
> can be replaced.

UUCP has already been replaced by NNCP: <http://www.nncpgo.org>.

"Why create yet another store-and-forward solution when UUCP, FTN and
even SMTP exists? Look in comparison section! Simplicity, cryptographic
security, sneakernet compatibility and easy integration with current
SMTP servers are the reasons."

What is NNCP? <https://www.complete.org/nncp>

"NNCP is to UUCP what ssh is to telnet; NNCP is an Encrypted,
authenticated, onion-routed version of UUCP!"

$ apt show nncp
NNCP is a package facilitating secure store-and-forward file and mail
exchange. It can be thought of as a modern UUCP with Internet smarts.

$ apt-get install nncp

$ man nncp
nncp - Node to Node Copy Program ...

Whether or not it is portable to C64 I don't know. But for anything
close to modern it replaces all UUCP use cases with wavy gravy.

--
K_|I_|N_|G_ 44...@king.bee | 38...@sugar.bug
B_|E_|E_|__ https://syfershock.com
S_|U_|G_|A_ news://alt.sources.crypto
R_|B_|U_|G_ #CipherTag #WaffleTag #Cryptologer

[Paul Edwards]

On 07/01/24 12:29, Ray Banana wrote:
> Thus spake Paul Edwards <muta...@gmail.com>
>
> [...]
>> ok, so I'm on Windows, not expecting to have
>> a cunbatch, nor the ability to run a shell script,
>> and 0D0A for line terminators is more normal too.
>
> Remember that both U's mean UNIX, so 0A is the standard line end.

Sure - I'm pointing out the differences
that I have on my system.

>> So - what are my options? 7z?
>
> In the source code you will find a file named rnews.c
>
> All batches should be processed by rnews, it knows how to cunbatch
> these input files.

Ok, thanks. So that's the way I would do it if I
were using more than just uucico.

But I'm only interested in uucico and I will be
replacing the rest over time.

So I eventually found this:

https://ncedc.org/qug/software/dos_utilities/comp430s.zip

Public domain C source code. I intend to bang
this into shape (suitable for my use on PDOS),
but so far I have already run it successfully
on my Linux system, and it has successfully
decompressed the above batch (leaving in the
rnews of course), after stripping off that
cunbatch line (and without adding a terminating
newline like vi did causing an error message
which took me a while to realize).

So this is actually good news for me. I don't
have any compression ability whatsoever in
PDOS currently (I do have unzip though - but
on UC386 I don't even have that).

So this traditional compress will give me an
option I didn't previously have.

It's unclear to me whether the Lempel-Ziv
algorithm used by the above code is also one
of the compression algorithms in info-zip.

If I could use this code as part of a "poor
man's zip", ie compatible with info unzip,
that would be fantastic.

Noting that old unzip was mostly public domain,
but zip has never been.

Anyway, back to me to do some programming work.
I now have raw data - thanks Ray!

BFN. Paul.

[J. P. Gilliver]

In message <und7h6$v8qv$1...@dont-email.me> at Sun, 7 Jan 2024 11:58:16,
Paul Edwards <muta...@gmail.com> writes
[]

>One day I may be last man standing.

[]
In which case, I can't see you'll have a use for _any_ communication
protocol.
(-:
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

I don't have an agree that our language torture is a quality add
- soldiersailor on Gransnet, 2018-3-8

[Paul Edwards]

On 07/01/24 12:01, immibis wrote:
> On 1/7/24 04:52, Paul Edwards wrote:
>> On 07/01/24 11:06, immibis wrote:
>>
>> I can't predict that the system I wish to connect
>> to has a C64-friendly protocol. UUCP is the standard.
>> And regardless - who is going to write this C64
>> protocol? Once again - UUCP actually exists.
>>
>
> NNTP is the standard, actually.

They both are. The C64 is neither a standard
nor even exists.

>>>> NEWNEWS presumably can't predict that you're
>>>> about to request the exact same newsgroups
>>>> that you got last time, and have a compressed
>>>> bundle waiting there for you.
>>>
>>> If you only have a 300 bps modem, you *really* don't want to transmit
>>> messages the recipient already has.
>>
>> I expect my peer system to only provide
>> new news via UUCP. That's the whole purpose
>> of it. New news, that my system has never
>> received before, all compressed and waiting
>> for me. The peer knows exactly what I have
>> and what I want - it provided it all to me.
>
> Do you expect to have more than one peer?

Not as I currently envisage it.

I'm planning for my daughter to be my peer.

I hide in the basement to prove it can be done,
and hand over a USB stick to do any communication
with the world.

At a later date I may move from my basement
to a Himalayan mountain to prove that works too.

Perhaps with a bunch of pigeons which I send
off once per week.

I do/did have communication with someone who
only connects to the internet after several
months or something. I don't know why. I wait
for him to provide the code changes to bwbasic.

Note that peering indirectly via my daughter
is technically a violation of the ES rules
(I believe), so at the point the software is
ready, I would have to negotiate with Ray or
go to a different server or come up with a
software solution that strips out that extra hop.

And it's a moot point if Europe has disappeared
in a nuclear war before the software is ready.
So I'll cross that bridge when I come to it.

And the bridge will be Fidonet, not Usenet, at
that point.

BFN. Paul.

[J. P. Gilliver]

In message <undcnv$vre6$1...@dont-email.me> at Sun, 7 Jan 2024 13:27:26,
Paul Edwards <muta...@gmail.com> writes
[]

>Perhaps with a bunch of pigeons which I send
>off once per week.

[]
You are of course familiar with RFCs 1149 and 2549. Though IPoAC and
CPIP may not suit your needs.

--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

resentment is like drinking poison and expecting the other person to die -
attributed to Carrie Fisher by Gareth McLean, in Radio Times 28 January-3
February 2012

[Paul Edwards]

On 07/01/24 12:00, immibis wrote:

>> So I wish to have the same setup with UUCP.
>> Yes, if you have a dialup available - great -
>> use it. If you have a 800 Mbps fiber-optic
>> link - even greater - use that instead. If
>> you don't, then sneakernet it.
>
> Presumably the peers you can reach through 800Mbps fiber-optic Internet
> aren't the same peers you can reach over sneakernet. You might be
> solving this problem one layer too low.

Oh - sorry - unstated assumption time.

I was intending to use Usenet to "prime" the
offline system.

So yeah, I encourage a programmer who lives
across town from me - or whose child attends
the same school as mine or whatever - to use
NNTP or whatever else he likes to access
comp.lang.c. As opposed to his interactive
Discord chatroom or whatever.

At least put some traffic into comp.lang.c,
or at least, read it.

Then when the Western powers are nuked out
of existence by Putin, I still exchange USB
sticks via the same local school. And we
spend a lot of time discussing if anyone
knows how to create a modem, since we still
have access to copper wire.

Something along those lines.

Basically I'm leery of relying on technology
I don't understand. It's too much for me
personally to know everything, but I do know
how to construct and maintain an operating
system. And over time I want to make sure
there is someone "local" who understands how
to build a modem. Things like that.

I haven't even started asking those questions.
Although I have the Fidonet technology ready
to go on USB sticks, I don't yet have the
ability to "prime" the system. The raw data
has started to come in though, and I now
understand it sufficiently to know how to
massage it for future use.

>>> You'll probably find UUCP spooling useful to hack such a system, because
>>> sneakernet is batch communication just like UUCP, but you won't actually
>>> use UUCP, just the parts that surround it.
>>
>> Sure.
>>
>> And as I've said previously, the actual
>> sneakernet will be Fidonet anyway.
>
> then I don't understand the discussion about UUCP over sneakernet.

I'm currently on Usenet. It's easier to discuss
Usenet than Fidonet. I will independently be
using FTN software. I mentioned before that my
use is designed to be transparent.

I will be using msged (fidonet), not Thunderbird.

But Ray would have no way of knowing that, other
than my odd use of X-Comment-To in practically
all of my messages. Because it is designed to
be read in Fidonet TOO.

BFN. Paul.

[Andy Burns]

Paul Edwards wrote:

> Just because I happen to be on a permanently-connected

> 800 Mbps fiber-optic link right now, I wish to be

> prepared for not just dialup, but sneakernet.

Usenet saves civilisation?

[Paul Edwards]

A Usenet/Fidonet combination.

I got my first modem in 1987 and C_ECHO I
think it was called, blew my mind away. I
had a technical question and got a answers
from different parts of the world (mainly
America). It may have taken a week for the
answers to come in - but I simply adjust
to that reality - and it's a hell of a lot
better reality than whatever the alternative
is (on your own with a mountain of books in
a library? you may as well just watch TV and
hope the answer comes up - there was a
"Peanuts" comic strip where that happens).

And there's not a lot of software/hardware
required to make this work - including even
the operating system.

Also - it's not necessarily after a catastrophe.

I remember reading about some remote place in
Cambodia without internet access.

And Bhutan struggles with internet too:

https://en.wikipedia.org/wiki/Sneakernet#Bhutan

And yes - Elon Musk is supposedly going to save
the world. If it happens, fine. If the Chinese
manage to destroy his satellites because it is
bypassing their firewall - maybe not so fine.
This gives you another option.

Note that even when I told fellow Australians
about how incredible Fidonet was - no-one was
interested. For some reason that morphed into
the entire world (including non-programmers)
permanently glued to their smartphones. I'm
not sure why they weren't doing as close as
they could get to that in 1987. It's not that
expensive, compared to e.g. a car - which they
all had, but I didn't.

So I read about people like Zuckerberg trying
to bring the Internet to remote areas of Africa
in impractical ways - but it seems that no
corporation has an interest in the hobbyist
Fidonet technology - with sneakernet substitute.
Not that surprising when there is nothing in it
for them.

But yeah - hobbyists bypassing the government
in 3rd world dictatorships sounds like an
interesting thing to me. The internet can and
is cut by those dictators. Very easily. Much
harder to stop people exchanging USB sticks.
Probably impossible to physically enforce. You
have to rely on threats of extreme punishment
instead. North Korea has something like that
going too. South Korean movies being exchanged
on USB stick.

What they're missing is a combination of Fidonet,
Usenet and sneakernet.

Maybe.

BFN. Paul.

[Jon Ribbens]

On 2024-01-07, Paul Edwards <muta...@gmail.com> wrote:

> And yes - Elon Musk is supposedly going to save the world.

Wat?

Even Elon Musk doesn't claim that Elon Musk is going to save the world,
as far as I'm aware.

[immibis]

On 1/7/24 05:56, SugarBug wrote:
> On Sun, 7 Jan 2024 11:52:38 +0800
> Paul Edwards <muta...@gmail.com> wrote:
>
>> I can't predict that the system I wish to connect
>> to has a C64-friendly protocol. UUCP is the standard.
>> And regardless - who is going to write this C64
>> protocol? Once again - UUCP actually exists.
>>
>> Yes, in the future - when someone volunteers to
>> write the software (crickets all round), UUCP
>> can be replaced.
>
> UUCP has already been replaced by NNCP: <http://www.nncpgo.org>.
>

Only in the eyes of the person who created NNCP. If you have UUCP over
TCP, and a TLS certificate, running UUCP over TLS is only a small change.

[Paul Edwards]

Other people keep "reassuring" me that there
is no need for the Fidonet technology I describe -
Elon Musk is going to make sure that everyone has
access to the internet.

I point out that the costs are expensive for people
in the 3rd world, but they assure me that the prices
will come down, so there is zero need for my low
tech solution. Internet has you covered, baby.

Then there is this:

https://en.wikipedia.org/wiki/One_Laptop_per_Child#Criticism

Mohammed Diop specifically criticized the project as an attempt to
exploit the governments of poor nations by making them pay for hundreds
of millions of machines and the need of further investments into
internet infrastructure


Note that my "Fidonet technology" can be
operated using smartphones too. I even paid
a Filipino to translate the instructions into
Tagalog. The Philippines doesn't seem to be
a good place to do it (unless a nuclear war
forces the issue as the ubiquitous Facebook
disappears). It doesn't seem to have the
issues of Bhutan, and I only have limited
insight into Africa. Last feedback I got
(from Nigeria - an English-speaking country)
was:

As we interacted with some of the school teachers, they raised concerns
about the inability of their wards to understand basic English. Although
they presented the brightest students, I hold some reservations about
their ability to grasp the learning required to fully operate PDNet in a
short while, seeing that these students have not been exposed to or
operated a computer system until now. It might be a huge leap.


I'm not sure what to do about that. Maybe
I need to move to Vietnamese jungles instead.

Or maybe it's a simple case of needing a
translation. Note that "PDNet" is what I
am calling my Fidonet-technology (more-or-less)
network.

BTW - you can attach an external keyboard
that costs about US$2 to a smartphone and
use it as a computer. Remember I mentioned
the C64? Maybe it's not so much a C64 as a
smartphone running Bochs. Note that I have
a native ARM version too, but it's still
proof of concept.

Bochs on a smartphone doesn't allow you to
communicate via the serial port. So you
need to get the mail bundles in manually.
You can't run a news server either.

BFN. Paul.

[Paul Edwards]

And the content I am posting/receiving is
public news messages. The only thing worth
encrypting is the password.

A password that would give you access to ...
public news messages which you can easily get.

And attempting to use my password will
potentially expose you. For what benefit?

Vandalism is the only possible benefit,
so it's not a zero risk. But exposing and
potentially jailing vandals has a benefit too.

Either way I'll cross that bridge when I get
to it - and it is not relevant either way when
I get to the jungles (or whatever they have)
of Bhutan.

BFN. Paul.

[Paul Edwards]

On 07/01/24 12:29, Ray Banana wrote:

> All batches should be processed by rnews, it knows how to cunbatch
> these input files.

BTW, is there a way of adding/deleting
newsgroups from my feed?

On Fidonet I would send an email to
areafix I think it was, and some software
would automatically make the appropriate
changes. Does UUCP have an equivalent?

I don't need any changes at the moment
(and I am reluctant to make changes in
the future if it involves annoying someone
in private).

Thanks. Paul.


P.S. I just did another pull - a bit less
than 24 hours. And now I have exceeded the
capacity of a 360k floppy disk in total.
So yeah - this works. I can remember being
frustrated at using someone else's floppy-only
XT computer.


2024-01-07 11:12a 2,417 6)gh0i1
2024-01-07 11:12a 2,947 6)gh0jh
2024-01-07 11:12a 2,142 6)gh0i2
2024-01-07 11:12a 4,376 6)gh0ji
2024-01-07 11:12a 1,826 6)gh0jj
2024-01-08 06:14a 3,964 6)gh0n4
2024-01-08 06:14a 10,667 6)gh0n0
2024-01-08 06:14a 3,131 6)gh0n1
2024-01-08 06:14a 2,860 6)gh0n2
2024-01-08 06:14a 4,033 6)gh0n3
2024-01-08 06:14a 1,480 6)gh0n5
2024-01-08 06:14a 4,084 6)gh0n6
2024-01-08 06:14a 3,974 6)gh0n7
2024-01-08 06:14a 4,457 6)gh0n8
2024-01-08 06:14a 4,273 6)gh0n9
2024-01-08 06:14a 2,425 6)gh0n@
2024-01-08 06:14a 2,251 6)gh0n^
2024-01-08 06:14a 1,378 6)gh0n_
2024-01-08 06:14a 4,133 6)gh0n`
2024-01-08 06:14a 2,272 6)gh0na
2024-01-08 06:14a 2,105 6)gh0o8
2024-01-08 06:14a 3,628 6)gh0o9
2024-01-08 06:14a 6,101 6)gh0o@
2024-01-08 06:14a 1,210 6)gh0o^
2024-01-08 06:14a 1,199 6)gh0o_
2024-01-08 06:14a 1,632 6)gh0o`
2024-01-08 06:14a 2,798 6)gh0oa
2024-01-08 06:14a 1,262 6)gh0ob
2024-01-08 06:14a 3,325 6)gh0ol
2024-01-08 06:14a 1,909 6)gh0oc
2024-01-08 06:14a 1,273 6)gh0od
2024-01-08 06:14a 2,752 6)gh0om
2024-01-08 06:14a 1,148 6)gh0on
2024-01-08 06:14a 3,052 6)gh0oo
2024-01-08 06:15a 1,716 6)gh0op
2024-01-08 06:15a 2,797 6)gh0oq
2024-01-08 06:15a <DIR> .
2024-01-08 06:15a 2,093 6)gh0or
2024-01-08 06:15a <DIR> ..
96 File(s) 393,337 bytes

[immibis]

(I previously wrote a message, then hit some keys and it disappeared.
I'm not sure whether it was sent or deleted. Apologies for any duplicates.)

On 1/7/24 07:47, Paul Edwards wrote:
> Then when the Western powers are nuked out
> of existence by Putin, I still exchange USB
> sticks via the same local school. And we
> spend a lot of time discussing if anyone
> knows how to create a modem, since we still
> have access to copper wire.

Why not a fiber connection to your local school? (Better yet, a
hackerspace. If you don't have one, start one. You could be your town's
network hub after the hypothetical nuke scenario.)

Worried about an EMP destroying all technology *except* for Commodore 64s?

Can you create a computer?

>
> Something along those lines.
>
> Basically I'm leery of relying on technology
> I don't understand. It's too much for me
> personally to know everything, but I do know
> how to construct and maintain an operating
> system. And over time I want to make sure
> there is someone "local" who understands how
> to build a modem. Things like that.
>

Sounds like a good excuse to start understanding it.

I repeat the recommendation for amateur radio as it seems like you might
enjoy it. Many hams build their own equipment, and understanding the
electronic principles is a requirement to get a licence. (They're still
debating whether to remove the Morse code knowledge requirement.)



You don't need a modem for a short-distance link (but you do for radio).
A modem converts a digital signal into one that meets the analog
specifications for a certain type of transmission channel. If you have
full control over the link, you can use a direct digital link - connect
two serial ports together. To prevent ground loops you want some kind of
isolator circuit in the middle, and you probably want lightning
protection at both ends, but those are implementation details. You can
also run serial port protocol directly over fiber-optic, which doesn't
have these problems. At the low speeds of a serial port, you can hook up
any random LED to the transmit line, any random phototransistor and a
bias resistor to the receiver line, and you could probably use Christmas
decorations as the cable, although real fiber-optic cable doesn't cost
much (30 cents per meter in low quantities).



You can also build a modem in software. Make a program that assembles a
stream of bits, then sends a high frequency for 1, and a low frequency
for 0. This is literally how the old hardware 3- and 4-digit modems worked.
You'll need to synchronize timing, which was taken care of at the serial
port layer in the old hardware modems. One simple way to do this is to
add a third sync frequency which is sent every certain number of bits,
and the receiver adjusts its bit timer to match that exact interval.
Another way is to reset the bit timer every time a bit transition is
received, and if too many of the same bit are sent in a row, insert an
escape sequence to force a transition. The serial port protocol (UART)
just inserts a 0 at the start of each byte and a 1 at the end, which
guarantees at least two transitions per byte.

> I haven't even started asking those questions.

The best time to start was 20 years ago. The second best time is now.

> Although I have the Fidonet technology ready
> to go on USB sticks, I don't yet have the
> ability to "prime" the system. The raw data
> has started to come in though, and I now
> understand it sufficiently to know how to
> massage it for future use.

Keep in mind: the primary purpose of Usenet (and Fidonet) is real-time
(by 1980 standards) communication. It isn't a static archive. You need
to remain linked to the system to participate. You can't disconnect and
still communicate with people.

If you want to communicate with far away people via Usenet, you need a
connection to them. If you don't, you have no need to "prime" anything.
If you want some resilient links to local people, it's better to set
them up now, not when a disaster happens.

> I'm currently on Usenet. It's easier to discuss
> Usenet than Fidonet. I will independently be
> using FTN software. I mentioned before that my
> use is designed to be transparent.
>
> I will be using msged (fidonet), not Thunderbird.
>
> But Ray would have no way of knowing that, other
> than my odd use of X-Comment-To in practically
> all of my messages. Because it is designed to
> be read in Fidonet TOO.

I think you are worrying too much about moderation. Rules seem to be
pretty loose on Usenet. The only rule that really seems to be enforced
is "don't spam." E-S has some more rules, like "don't impersonate
people", that will be enforced only if they need to be. Don't create a
problem, and you won't be subject to enforcement action.

I read that Fidonet has a hierarchical, authoritarian administration.
Usenet is relative anarchy. If you can convince someone to link their
server to your own (which is not hard - check out net.admin.peering) you
can make your server part of Usenet. You don't need any permission from
any central authority.

E-S isn't the only Usenet server by far, though I think it's the biggest
non-binary public one, besides Google Groups. Many people seem to post
from their own niche servers - though obviously not in eternal-september
groups.

[immibis]

Reminds me of "you can't solve social problems with technical
solutions". Apparently you can't create social problems with technology,
either. What made everyone glued to phones was *not* simply the
technology of phones - it's much more complicated than that. Otherwise
we'd have seen everyone glued to their Blackberry, but we didn't see
that - it started with the iPhone, possibly because of Steve Jobs's
charisma.

Recently I've seen people talking about blue bubbles and green bubbles.
iPhone users see chat messages in "bubbles" (the visual box surrounding
each message) and they see messages from other iPhone users in blue
bubbles, and non-iPhone users in green bubbles. Apparently, having blue
bubbles is a social status signal to all your friends, so much that some
non-iPhone-users use third-party apps transmit all their messages
through third-party servers and the iPhone messaging system just to show
blue bubbles to their friends.

It's ridiculous, and it's also a purely social thing, not a
technological thing. If non-Apple phones somehow had the native ability
to use iMessages (the EU might require this) it wouldn't confer the same
positive social signal - instead, the signal would lose its social value
and be replaced by something else.



The Internet is also like this. We aren't all Internet addicts because
the Internet exists, we're Internet addicts because it's socially
expected. People use their phones for banking not just because they can,
but also because banks *require* it. If you're required to have a phone
to have a bank account, a job, a house, and so on, you may as well use
it for "silly" things like looking up facts on Wikipedia, or listening
to podcasts, or posting selfies of your food.

A certain percentage of people are naturally going to pick up computery,
Internetty hobbies, just like a percentage pick up archery, or go-kart
racing, or swimming, or underwater basket weaving. Push computers and
Internet into people's faces constantly, and a lot more of them will
pick them up. And there's a feedback loop, too - those other areas have
less people and fall into decline, leading to less people "naturally"
picking them up, and more people in the areas that didn't decline.

>
> So I read about people like Zuckerberg trying
> to bring the Internet to remote areas of Africa
> in impractical ways - but it seems that no
> corporation has an interest in the hobbyist
> Fidonet technology - with sneakernet substitute.
> Not that surprising when there is nothing in it
> for them.
>
> But yeah - hobbyists bypassing the government
> in 3rd world dictatorships sounds like an
> interesting thing to me. The internet can and
> is cut by those dictators. Very easily. Much
> harder to stop people exchanging USB sticks.
> Probably impossible to physically enforce. You
> have to rely on threats of extreme punishment
> instead. North Korea has something like that
> going too. South Korean movies being exchanged
> on USB stick.
>

It does sound interesting, and it is interesting, but if me or you
actually lived in that scenario, we'd have much bigger problems than a
lack of communication. If you live in North Korea, how will
communicating with the outside world make your life better? It's more
likely to get you killed. At best, you learn that other people have
better lives, but you already knew that.

I'll always remember that some organization like the EFF was promoting
Tor to enable the Uyghurs in China (who are being genocided) to
communicate without tracking. A Uyghur responded: "You don't understand.
*They're living in my house.*"
What use are technological workarounds to them? If they play with
technology, they get killed.

[immibis]

On 1/7/24 22:15, Paul Edwards wrote:

> [snip]

I'd also like to point out that Fidonet, Usenet and sneakernet aren't
equivalents. Fidonet and Usenet are whole systems, but sneakernet is a
point-to-point data transfer (along the same level as RS-232).

You can transfer Usenet traffic or (presumably) Fidonet traffic, as well
as other types of traffic, like UUCP bang-path email, through a sneakernet.

[J. P. Gilliver]

In message <unf88q$17ol1$1...@dont-email.me> at Sun, 7 Jan 2024 23:23:21,
immibis <ne...@immibis.com> writes
[]

>I repeat the recommendation for amateur radio as it seems like you
>might enjoy it. Many hams build their own equipment, and understanding
>the electronic principles is a requirement to get a licence. (They're
>still debating whether to remove the Morse code knowledge requirement.)
>

FWIW, the Morse requirement is mostly if not entirely gone from the UK
amateur licence. It started to go (I think) in the late 1970s: when I
got mine around 1980, it wasn't needed to become a holder of an "Amateur
Radio Licence B" (often referred to by hams as a "class B", but the
authorities never used the word "class" in its title); at that time and
for some decades afterwards, B licencees were restricted to VHF and
above (bands above 100 MHz). Some decades later, B licencees were
allowed access to HF, i. e. the same bands as A licencees (those who'd
passed the Morse test); I don't know if there's _anything_ that As can
still do that Bs cannot, i. e. I'm not sure if passing the Morse gets
you _anything_ extra these days, or even if the test is still
administered. (They also introduced a more "beginner" type of licence in
the meantime - they're restricted to fewer bands, powers, and modes than
the A and B licence holders.)
[]

>You can also build a modem in software. Make a program that assembles a
>stream of bits, then sends a high frequency for 1, and a low frequency
>for 0. This is literally how the old hardware 3- and 4-digit modems
>worked.
>You'll need to synchronize timing, which was taken care of at the
>serial port layer in the old hardware modems. One simple way to do this
>is to add a third sync frequency which is sent every certain number of
>bits, and the receiver adjusts its bit timer to match that exact
>interval. Another way is to reset the bit timer every time a bit
>transition is received, and if too many of the same bit are sent in a
>row, insert an escape sequence to force a transition. The serial port

(Known as "bit stuffing". [A more sophisticated version of the same
thing is used in the data stream for burning CDs.])
[]

>E-S isn't the only Usenet server by far, though I think it's the
>biggest non-binary public one, besides Google Groups. Many people seem
>to post from their own niche servers - though obviously not in
>eternal-september groups.
>

I assumed that too, but apparently at least some of the e-s 'groups
_are_ propagated outside e-s. Presumably like the old demon.* 'groups,
and probably a few other such sets.

>
>
>
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

I hope you dream a pig.

[Paul Edwards]

Sorry - I don't really understand this.

They all look the same to me. You exchange
messages with someone.

I'm only talking about public newsgroups/echoes,
not email/netmail.

Currently (well, over in the other window), I
am receiving UUCP batches. Sounds point to point
to me. Ray could literally give me a USB stick
instead. He could in fact put a 360k floppy disk
in the postal mail every day. I bet if I paid him
the right amount of money he would do exactly that.

I'll answer your other (great) messages a bit later.

BFN. Paul.

[immibis]

On 1/8/24 00:43, Paul Edwards wrote:
> On 08/01/24 06:41, immibis wrote:
>> On 1/7/24 22:15, Paul Edwards wrote:
>>> [snip]
>>
>> I'd also like to point out that Fidonet, Usenet and sneakernet aren't
>> equivalents. Fidonet and Usenet are whole systems, but sneakernet is a
>> point-to-point data transfer (along the same level as RS-232).
>>
>> You can transfer Usenet traffic or (presumably) Fidonet traffic, as well
>> as other types of traffic, like UUCP bang-path email, through a
>> sneakernet.
>
> Sorry - I don't really understand this.
>
> They all look the same to me. You exchange
> messages with someone.
>
> I'm only talking about public newsgroups/echoes,
> not email/netmail.

Isn't that a big difference between Fidonet and Usenet - that Fidonet
also supports directed emails?

>
> Currently (well, over in the other window), I
> am receiving UUCP batches. Sounds point to point
> to me.

Right, UUCP, not Usenet. You're running Usenet over UUCP, you're talking
about Usenet over sneakernet, and you're presumably posting to Usenet
over NNTP since your replies aren't a day late.

At some point you wrote something that put Usenet, Fidonet, and
sneakernet at the same level, which is incorrect. Sneakernet is more on
the level of UUCP.

[Paul Edwards]

On 08/01/24 08:29, immibis wrote:

>>> I'd also like to point out that Fidonet, Usenet and sneakernet aren't
>>> equivalents. Fidonet and Usenet are whole systems, but sneakernet is a
>>> point-to-point data transfer (along the same level as RS-232).
>>>
>>> You can transfer Usenet traffic or (presumably) Fidonet traffic, as well
>>> as other types of traffic, like UUCP bang-path email, through a
>>> sneakernet.
>>
>> Sorry - I don't really understand this.
>>
>> They all look the same to me. You exchange
>> messages with someone.
>>
>> I'm only talking about public newsgroups/echoes,
>> not email/netmail.
>
> Isn't that a big difference between Fidonet and Usenet - that Fidonet
> also supports directed emails?

Well, Usenet - and ES in particular - requires you
to put an email address in all of your messages, so
to me it is "the internet" or something (Usenet +
email) and the equivalent of Fidonet. I'm not
familiar with how email works. I don't know much
about the internet either. I'm not really
interested either - I'm after the simpler technology.

Even if it is emulated/simulated, e.g. getting
qemu to direct the serial port to TCP/IP.

>> Currently (well, over in the other window), I
>> am receiving UUCP batches. Sounds point to point
>> to me.
>
> Right, UUCP, not Usenet. You're running Usenet over UUCP, you're talking
> about Usenet over sneakernet, and you're presumably posting to Usenet
> over NNTP since your replies aren't a day late.

I see.

BTW, it is true that I'm using NNTP via Thunderbird,
but if I were using UUCP there would not be a delay
either. I would poll ES whenever I had free time to
read/write messages. The only (likely) difference is
that I would probably delay sending until I had
finished the current batch, so suddenly you would
receive 3 messages at the same time.

> At some point you wrote something that put Usenet, Fidonet, and
> sneakernet at the same level, which is incorrect. Sneakernet is more on
> the level of UUCP.

Sure.

BFN. Paul.

[Paul Edwards]

On 08/01/24 06:23, immibis wrote:

>> Then when the Western powers are nuked out
>> of existence by Putin, I still exchange USB
>> sticks via the same local school. And we
>> spend a lot of time discussing if anyone
>> knows how to create a modem, since we still
>> have access to copper wire.
>
> Why not a fiber connection to your local school? (Better yet, a
> hackerspace. If you don't have one, start one. You could be your town's
> network hub after the hypothetical nuke scenario.)

I'm not sure that is allowed, but even if it
was, that would be real money spent on something
I don't think is actually going to happen.

What I am currently doing is just a hobby for
a vague and theoretical situation.

Only time is spent on writing the software, or
other associated things (including this entire
thread).

> Worried about an EMP destroying all technology *except* for Commodore 64s?

No. In a theoretical calamity (e.g. Covid-24
wiping out 99.9% of the world's population),
and after a couple of centuries/millenia of
recovery, any new computers being built would
start with being 8-bit - at least possibly.

And I want to have the PDOS source code
punched on plastic punch cards so that it
can be married up to the new computers.

Well. I don't think PDOS will ever fit in
64k, but when 16-bit segmented or 32-bit
flat computers are finally around, to give
me 2 MiB of memory, then PDOS can be used.

> Can you create a computer?

Nope. But someone sent me a link showing
that you can construct a CPU with just
ROM chips.

>> Basically I'm leery of relying on technology
>> I don't understand. It's too much for me
>> personally to know everything, but I do know
>> how to construct and maintain an operating
>> system. And over time I want to make sure
>> there is someone "local" who understands how
>> to build a modem. Things like that.
>
> Sounds like a good excuse to start understanding it.

I'll be dead before too long, so I'm more
interested in making sure that my daughter
and her friends know how to maintain the
technology they use.

Have you seen the movie "Idiocracy"?

> I repeat the recommendation for amateur radio as it seems like you might
> enjoy it. Many hams build their own equipment, and understanding the
> electronic principles is a requirement to get a licence. (They're still
> debating whether to remove the Morse code knowledge requirement.)

I'm already enjoying what I do, and I'm already
flat out just on the software side.

In the last 24 hours I have found public domain
"compress" and also public domain "zip" (hwzip).
Those are welcome additions to UC386 (a pure
PD distribution) which currently has no
compression capability at all.

> You don't need a modem for a short-distance link (but you do for radio).
> A modem converts a digital signal into one that meets the analog
> specifications for a certain type of transmission channel. If you have
> full control over the link, you can use a direct digital link - connect
> two serial ports together. To prevent ground loops you want some kind of
> isolator circuit in the middle, and you probably want lightning
> protection at both ends, but those are implementation details. You can
> also run serial port protocol directly over fiber-optic, which doesn't
> have these problems. At the low speeds of a serial port, you can hook up
> any random LED to the transmit line, any random phototransistor and a
> bias resistor to the receiver line, and you could probably use Christmas
> decorations as the cable, although real fiber-optic cable doesn't cost
> much (30 cents per meter in low quantities).

That certainly sounds cool.

But so does a UUCP batches to FTN convertor,
as that is needed in either case.

> escape sequence to force a transition. The serial port protocol (UART)
> just inserts a 0 at the start of each byte and a 1 at the end, which
> guarantees at least two transitions per byte.

Ok. As someone else mentioned - CDs
have a code for a better version of that
(Solomon-Reed).

>> I haven't even started asking those questions.
>
> The best time to start was 20 years ago. The second best time is now.

I would argue the best time is after I have the
source code ready to use the hardware, as that
utilizes my existing skills.

And the hardware requirements in the event of
the nuclear war (which could be 600 years from
now), may have changed. Maybe copper is cheaper
than glass or we have room temperature
superconductors or whatever.

>> Although I have the Fidonet technology ready
>> to go on USB sticks, I don't yet have the
>> ability to "prime" the system. The raw data
>> has started to come in though, and I now
>> understand it sufficiently to know how to
>> massage it for future use.
>
> Keep in mind: the primary purpose of Usenet (and Fidonet) is real-time
> (by 1980 standards) communication. It isn't a static archive. You need
> to remain linked to the system to participate. You can't disconnect and
> still communicate with people.
>
> If you want to communicate with far away people via Usenet, you need a
> connection to them. If you don't, you have no need to "prime" anything.
> If you want some resilient links to local people, it's better to set
> them up now, not when a disaster happens.

To get a critical mass of people to talk to,
I need Usenet and an international audience.

If I can have 1 other person in my local area
join that (this) international conversation,
then I am prepared for when the rest of the
world disappears (could be due to sanctions
rather than nuclear war).

>> I'm currently on Usenet. It's easier to discuss
>> Usenet than Fidonet. I will independently be
>> using FTN software. I mentioned before that my
>> use is designed to be transparent.
>>
>> I will be using msged (fidonet), not Thunderbird.
>>
>> But Ray would have no way of knowing that, other
>> than my odd use of X-Comment-To in practically
>> all of my messages. Because it is designed to
>> be read in Fidonet TOO.
>
> I think you are worrying too much about moderation. Rules seem to be
> pretty loose on Usenet. The only rule that really seems to be enforced
> is "don't spam." E-S has some more rules, like "don't impersonate
> people", that will be enforced only if they need to be. Don't create a
> problem, and you won't be subject to enforcement action.

It's more than just impersonation.

You are required to be the only person sending
messages. You can't be a gateway for others.

> I read that Fidonet has a hierarchical, authoritarian administration.
> Usenet is relative anarchy. If you can convince someone to link their
> server to your own (which is not hard - check out net.admin.peering) you
> can make your server part of Usenet. You don't need any permission from
> any central authority.

Sure. And PDNet operates under the same
principles as far as possible. This will
cause a clash of numbers though, so you
just have to cooperate to avoid that. Or
ignore it - up to you.

BFN. Paul.


(BTW, electricity is out right now in my
area of the Philippines - it often goes
out - so that's another thing I wish to
prepare for - offline reading/writing
due to genuine no-internet conditions).

[Paul Edwards]

On 08/01/24 06:38, immibis wrote:

>> Note that even when I told fellow Australians
>> about how incredible Fidonet was - no-one was
>> interested. For some reason that morphed into
>> the entire world (including non-programmers)
>> permanently glued to their smartphones. I'm
>> not sure why they weren't doing as close as
>> they could get to that in 1987. It's not that
>> expensive, compared to e.g. a car - which they
>> all had, but I didn't.
>
> Reminds me of "you can't solve social problems with technical
> solutions". Apparently you can't create social problems with technology,
> either. What made everyone glued to phones was *not* simply the
> technology of phones - it's much more complicated than that. Otherwise
> we'd have seen everyone glued to their Blackberry, but we didn't see
> that - it started with the iPhone, possibly because of Steve Jobs's
> charisma.

That's an interesting take, but I don't think
that is what is causing it.

> Recently I've seen people talking about blue bubbles and green bubbles.
> iPhone users see chat messages in "bubbles" (the visual box surrounding
> each message) and they see messages from other iPhone users in blue
> bubbles, and non-iPhone users in green bubbles. Apparently, having blue
> bubbles is a social status signal to all your friends, so much that some
> non-iPhone-users use third-party apps transmit all their messages
> through third-party servers and the iPhone messaging system just to show
> blue bubbles to their friends.

Here in the Philippines (noting that I am
from Australia), I don't think people are
showing off their phones for social status.
It's all-Android here btw. No-one is interested
in your phone. They are only interested in their
own phone. To use. Not to show you.

> It's ridiculous, and it's also a purely social thing, not a
> technological thing. If non-Apple phones somehow had the native ability
> to use iMessages (the EU might require this) it wouldn't confer the same
> positive social signal - instead, the signal would lose its social value
> and be replaced by something else.

I don't remember anyone in Australia showing
off their iPhone to me either - to impress
me or something like that. They might ask
"do you know how to do xyz on an iphone?".

> The Internet is also like this. We aren't all Internet addicts because
> the Internet exists, we're Internet addicts because it's socially
> expected.

I don't see anyone (individual) expecting you
to use a smartphone.

If anyone fits that description it is me. For
the same reason I was trying to push modems on
people in 1987. I'm the computer equivalent of
a Christian evangelist. I quite literally hand
out free phones and computers to people, instead
of bibles.

I would potentially give you a (cheap)
computer or more likely smartphone (to
be used as a computer) if it was going
to be used to run PDOS/get PDNet going
or similar.

> People use their phones for banking not just because they can,
> but also because banks *require* it. If you're required to have a phone
> to have a bank account, a job, a house, and so on, you may as well use
> it for "silly" things like looking up facts on Wikipedia, or listening
> to podcasts, or posting selfies of your food.

Most people here don't have bank accounts.
In part because there's a maintaining balance
that they aren't willing to maintain. About US$40.

I see perfectly intelligent hardworking people
going to the bank instead of doing it online.
They have smartphones. They do post pictures in
Facebook group chat.

I don't see that in Australia though.

> A certain percentage of people are naturally going to pick up computery,
> Internetty hobbies, just like a percentage pick up archery, or go-kart
> racing, or swimming, or underwater basket weaving. Push computers and
> Internet into people's faces constantly, and a lot more of them will
> pick them up. And there's a feedback loop, too - those other areas have
> less people and fall into decline, leading to less people "naturally"
> picking them up, and more people in the areas that didn't decline.

Ok.

>> But yeah - hobbyists bypassing the government
>> in 3rd world dictatorships sounds like an
>> interesting thing to me. The internet can and
>> is cut by those dictators. Very easily. Much
>> harder to stop people exchanging USB sticks.
>> Probably impossible to physically enforce. You
>> have to rely on threats of extreme punishment
>> instead. North Korea has something like that
>> going too. South Korean movies being exchanged
>> on USB stick.
>
> It does sound interesting, and it is interesting, but if me or you
> actually lived in that scenario, we'd have much bigger problems than a
> lack of communication. If you live in North Korea, how will
> communicating with the outside world make your life better?

The intention is not to communicate with the
outside world. That's what the rest of these
people are doing, because then they get a cut
of the action. Since I'm not after a cut of
the action (or at least - not in the same way),
I have/can afford a different approach.

> It's more likely to get you killed.

They're already risking that - to watch movies.

> I'll always remember that some organization like the EFF was promoting
> Tor to enable the Uyghurs in China (who are being genocided) to
> communicate without tracking. A Uyghur responded: "You don't understand.
> *They're living in my house.*"
> What use are technological workarounds to them? If they play with
> technology, they get killed.

I'm not sure I understand that. Who is living
in their house? The people they wish to
communicate with?

In my case, it is for North Koreans etc to
communicate a little further than their house.

A USB stick (already being passed around with
movies on it) to their neighbor or relative in
*that other house*.

With conversations (all in Korean - but using
the Korean alphabet, not their DBCS) along the
lines of:

Hey - I've noticed that all these communist
officials seem to be living high off the hog -
I thought communism was supposed to make
everyone equal?

Yeah - you're right - let's start a
non-communist revolution this time.


Or perhaps (all in Bhutanese or whatever they
speak in Bhutan - definitely a SBCS) - I thought
I read somewhere that they have a principle,
can't remember the name, let's say xxx, which
is this radical concept that sets their culture
above all others because it says that actions
should benefit all society, not just the
individual - and anyway, they could have a
conversation like:

Hey, if we have this xxx principle, why is it
that xxx doesn't extend to the dictator king -
why doesn't he care what we think and allow
us to be in charge?

Yeah, maybe we need a non-xxx revolution if
this is what xxx looks like in practice.


Note that I don't actually want a physical
revolution - 100,000 dead Iraqis in 1991
showed what happens when civilians armed
with pea-shooters go up against helicopter
gunships - I just want free-flowing debate
so that the hypocrisy is exposed and when
the time is right, people don't have
fantasies about the good old days when we
had Utopia.

Although with the Russians praising Stalin,
I don't hold out a lot of hope.

The Thais are the sort of people who can see
reality, but they don't need my software to
communicate.

So it's the right, theoretical, combination.

BFN. Paul.

[immibis]

On 1/8/24 04:49, Paul Edwards wrote:
>>>
>>> They all look the same to me. You exchange
>>> messages with someone.
>>>
>>> I'm only talking about public newsgroups/echoes,
>>> not email/netmail.
>>
>> Isn't that a big difference between Fidonet and Usenet - that Fidonet
>> also supports directed emails?
>
> Well, Usenet - and ES in particular - requires you
> to put an email address in all of your messages, so

E-S in particular allows, but recommends against, using a fake email
address ending in the reserved TLD (top-level domain) .invalid - for
example pa...@edwards.invalid - I've seen some people using these addresses.

You could also use the address of an email to Fidonet gateway combined
with your Fidonet address, if that's something which exists. The primary
concern is that you don't use someone else's address.

> to me it is "the internet" or something (Usenet +
> email) and the equivalent of Fidonet. I'm not
> familiar with how email works. I don't know much
> about the internet either. I'm not really
> interested either - I'm after the simpler technology.
>
> Even if it is emulated/simulated, e.g. getting
> qemu to direct the serial port to TCP/IP.
>

That's fine, but it would serve you well to also know the basic
principles behind the internet. I promise they aren't too complicated.
It's the implementation details that are complicated. Especially when it
comes to email techniques for preventing spam - that's a mess.

Perhaps the biggest conceptual difference between the internet and
whatever you're used to is that the internet layer is optimized to get
any packet anywhere in milliseconds. Most applications use a protocol
called TCP, which simulates something like a high-latency serial
connection between two internet hosts. Most applications establish a TCP
connection with whatever other host they wish to communicate with, then
treat it as if it was a direct connection. A TCP connection is addressed
not just to a certain host, but also to a certain process running on
that host.

There's no multicast or broadcast feature on the internet, so
applications that need such a feature have to implement it at the
application level, just like Usenet does.

> BTW, it is true that I'm using NNTP via Thunderbird,
> but if I were using UUCP there would not be a delay
> either. I would poll ES whenever I had free time to
> read/write messages. The only (likely) difference is
> that I would probably delay sending until I had
> finished the current batch, so suddenly you would
> receive 3 messages at the same time.
>

I see.

Of course, either UUCP or NNTP, or sneakernet, can be used in this way,
but newsreaders have no reason to artificially wait to submit messages
if they're able to connect to the server immediately.

[immibis]

On 1/8/24 06:19, Paul Edwards wrote:
>> escape sequence to force a transition. The serial port protocol (UART)
>> just inserts a 0 at the start of each byte and a 1 at the end, which
>> guarantees at least two transitions per byte.
>
> Ok. As someone else mentioned - CDs
> have a code for a better version of that
> (Solomon-Reed).
>

Actually that is for error correction. I wouldn't be surprised if it
forced transitions as a side effect, though.

Fast communication protocols typically use something called 8b/10b
encoding, which goes even farther by guaranteeing the number of 1s and
0s is exactly equal (to within a tolerance of +/- 1 of each). Even
faster ones use something called 64b/66b encoding, which only loosely
approximates it and apparently works well enough anyway.

> And the hardware requirements in the event of
> the nuclear war (which could be 600 years from
> now), may have changed. Maybe copper is cheaper
> than glass or we have room temperature
> superconductors or whatever.

Exactly. No point planning for a nuclear war 600 years in the future.

> It's more than just impersonation.
>
> You are required to be the only person sending
> messages. You can't be a gateway for others.

I can't find this rule. The closest is a rule saying you shouldn't share
your password.

[immibis]

On 1/8/24 06:19, Paul Edwards wrote:

> On 08/01/24 06:38, immibis wrote:
>
>>> Note that even when I told fellow Australians
>>> about how incredible Fidonet was - no-one was
>>> interested. For some reason that morphed into
>>> the entire world (including non-programmers)
>>> permanently glued to their smartphones. I'm
>>> not sure why they weren't doing as close as
>>> they could get to that in 1987. It's not that
>>> expensive, compared to e.g. a car - which they
>>> all had, but I didn't.
>>
>> Reminds me of "you can't solve social problems with technical
>> solutions". Apparently you can't create social problems with technology,
>> either. What made everyone glued to phones was *not* simply the
>> technology of phones - it's much more complicated than that. Otherwise
>> we'd have seen everyone glued to their Blackberry, but we didn't see
>> that - it started with the iPhone, possibly because of Steve Jobs's
>> charisma.
>
> That's an interesting take, but I don't think
> that is what is causing it.
>

Well, it was something about psychology and marketing, not something
about technology. I don't know what it was. If someone does, let me know.

My mum tried to push me to use Facebook a lot, and I never did. Why did
Facebook feel pointless to me but vital to her? I have no idea. Probably
the same reason I used to use Reddit a lot, before getting banned.

Probably the same reason I know a lot about computers instead of a lot
about fishing - chance. But chance influenced by what? I don't know.
Clearly something.

>> Recently I've seen people talking about blue bubbles and green bubbles.
>> iPhone users see chat messages in "bubbles" (the visual box surrounding
>> each message) and they see messages from other iPhone users in blue
>> bubbles, and non-iPhone users in green bubbles. Apparently, having blue
>> bubbles is a social status signal to all your friends, so much that some
>> non-iPhone-users use third-party apps transmit all their messages
>> through third-party servers and the iPhone messaging system just to show
>> blue bubbles to their friends.
>
> Here in the Philippines (noting that I am
> from Australia), I don't think people are
> showing off their phones for social status.
> It's all-Android here btw. No-one is interested
> in your phone. They are only interested in their
> own phone. To use. Not to show you.
>

I think it's more of an American thing. I'm a Kiwi living in Germany,
and nobody's asked me about my phone. Most people use Whatsapp or Signal.

> I don't see anyone (individual) expecting you
> to use a smartphone.

I need one to access my bank account. I need one to get an appointment
at the immigration office. I need one to get notifications about my
workplace. A desktop computer could substitute 2/3 of these, but that's
sacrificing convenient portability for absolutely no reason. It's good
for development and other work though.

Maybe in the near future I'll write an analogue of Quasseldroid for
Usenet, so I can access it on my phone.

> If anyone fits that description it is me. For
> the same reason I was trying to push modems on
> people in 1987. I'm the computer equivalent of
> a Christian evangelist. I quite literally hand
> out free phones and computers to people, instead
> of bibles.

Steve Jobs was also an evangelist, but for some reason, his version stuck.

Understanding and exploiting what makes things stick in a large
population is literally what makes people into technology billionaires.

But I still think it's mostly chance.

> I would potentially give you a (cheap)
> computer or more likely smartphone (to
> be used as a computer) if it was going
> to be used to run PDOS/get PDNet going
> or similar.

That would be an impedance mismatch. User interfaces designed for the
IBM PC's big screen and keyboard just don't work well on phones, because
they are fundamentally different devices. That's not to say their CPUs
don't work the same way, but a computer is more than just a CPU. The
human interface peripherals are radically different.

Fling-scrolling is very good interface design on a touchscreen but makes
no sense on a mouse (not counting mice with unlocked scroll wheels that
you can also fling).

Typing on a full-size keyboard is very good, once you have enough
practice, but typing on a phone-sized virtual keyboard on a touchscreen
is terrible due to the missing physical feedback.

Almost everything you do on a phone should be modal and use the entire
screen, because there isn't much screen space to use.

To select a file or folder on a phone, you should scroll through a list
by dragging or flinging it, and tap the item you want. Typing the name
is terrible UI design for a phone.

>> People use their phones for banking not just because they can,
>> but also because banks *require* it. If you're required to have a phone
>> to have a bank account, a job, a house, and so on, you may as well use
>> it for "silly" things like looking up facts on Wikipedia, or listening
>> to podcasts, or posting selfies of your food.
>
> Most people here don't have bank accounts.
> In part because there's a maintaining balance
> that they aren't willing to maintain. About US$40.

And there you see how it's a kind of social inertia almost totally
unrelated to technology. *My* bank *strictly requires* the use of a
smartphone. If I couldn't get one, I'd have to close my account. There
are other banks available that prefer to operate by post - even worse
idea IMO.

> I see perfectly intelligent hardworking people
> going to the bank instead of doing it online.
> They have smartphones. They do post pictures in
> Facebook group chat.

Note that you can't get cash out of your smartphone. You still have to
go to an ATM or a bank branch or a store for that.

>> It does sound interesting, and it is interesting, but if me or you
>> actually lived in that scenario, we'd have much bigger problems than a
>> lack of communication. If you live in North Korea, how will
>> communicating with the outside world make your life better?
>
> The intention is not to communicate with the
> outside world. That's what the rest of these
> people are doing, because then they get a cut
> of the action. Since I'm not after a cut of
> the action (or at least - not in the same way),
> I have/can afford a different approach.

Then who are you trying to communicate with?

>
>> It's more likely to get you killed.
>
> They're already risking that - to watch movies.

> I'm not sure I understand that. Who is living
> in their house? The people they wish to
> communicate with?

The people who are oppressing them.

Tor is designed to help people communicate without being watched, right?
So, some people were talking about how to get Tor to the Chinese
Uyghurs, so they could communicate without being watched.

But it's no use, because a computer can't protect the privacy from the
people who are *physically* watching you because the government has
assigned them to live in your house and make sure you don't do anything
the government doesn't like.

Technological solutions can't solve social or political problems.

>
> In my case, it is for North Koreans etc to
> communicate a little further than their house.
>
> A USB stick (already being passed around with
> movies on it) to their neighbor or relative in
> *that other house*.
>

Alright, that's fair. I don't know how you'd get the software to them,
though. Have you seen people's reviews of the leaked copy of Red Star
OS, which is the only OS they're allowed to use?

It marks files with the identifier of the machine that created them, and
it scans files for subversive content so the police can be sent to their
house.

> With conversations (all in Korean - but using
> the Korean alphabet, not their DBCS) along the
> lines of:
>
> Hey - I've noticed that all these communist
> officials seem to be living high off the hog -
> I thought communism was supposed to make
> everyone equal?

Remember, the person you pass the USB stick to will be *handsomely
rewarded* if they can tell the police who wrote this message.

That's how most political power structures maintain themselves, actually
- by rewarding people for oppressing each other, the people collectively
ensure their own oppression.

> Note that I don't actually want a physical
> revolution - 100,000 dead Iraqis in 1991
> showed what happens when civilians armed
> with pea-shooters go up against helicopter
> gunships - I just want free-flowing debate
> so that the hypocrisy is exposed

what use is that? identifying hypocrisy doesn't protect you from the
guns the hypocrites are wielding.

[David W. Hodgins]

On Sun, 07 Jan 2024 22:49:39 -0500, Paul Edwards <muta...@gmail.com> wrote:
> BTW, it is true that I'm using NNTP via Thunderbird,
> but if I were using UUCP there would not be a delay
> either. I would poll ES whenever I had free time to
> read/write messages. The only (likely) difference is
> that I would probably delay sending until I had
> finished the current batch, so suddenly you would
> receive 3 messages at the same time.

You may want to look into the package leafnode.
https://www.leafnode.org/

Regards, Dave Hodgins

[Marco Moock]

Am 08.01.2024 um 06:04:03 Uhr schrieb Paul Edwards:

> A password that would give you access to ...
> public news messages which you can easily get.
>
> And attempting to use my password will
> potentially expose you. For what benefit?

A small one, but if implementing security against it is rather easy,
why don't do it?

> Vandalism is the only possible benefit,
> so it's not a zero risk. But exposing and
> potentially jailing vandals has a benefit too.

Mostly impossible. In how many countries spam is prohibited?
How many of the spammers are in jail?

[J. P. Gilliver]

In message <ung63q$1eti3$1...@dont-email.me> at Mon, 8 Jan 2024 07:52:42,
immibis <ne...@immibis.com> writes

>On 1/8/24 06:19, Paul Edwards wrote:
>>> escape sequence to force a transition. The serial port protocol (UART)
>>> just inserts a 0 at the start of each byte and a 1 at the end, which
>>> guarantees at least two transitions per byte.
>> Ok. As someone else mentioned - CDs
>> have a code for a better version of that
>> (Solomon-Reed).
>>
>
>Actually that is for error correction. I wouldn't be surprised if it
>forced transitions as a side effect, though.
>
>Fast communication protocols typically use something called 8b/10b
>encoding, which goes even farther by guaranteeing the number of 1s and
>0s is exactly equal (to within a tolerance of +/- 1 of each). Even
>faster ones use something called 64b/66b encoding, which only loosely
>approximates it and apparently works well enough anyway.

[]
There are two low-level requirements (error-correction is at a higher
level) that are the reason for bit-stuffing: clock recovery, and DC
removal/AC coupling.

Clock recovery: when two completely independent entities are
communicating over a serial line, the receiving one has to know when to
sample that line to determine whether what's coming in is a 1 or a 0. To
do this, it runs an oscillator - in software or hardware, it doesn't
matter - that counts down from the beginning of each bit, and samples at
the half way point; if you're sending bits 100 milliseconds long (10
bits per second), then the receiving end ideally samples at 50 ms into
each bit. It gets reset by the incoming transitions, i. e. when the
incoming signal switches from 0 to 1 or vice versa. If the sending end
sends tens (or hundreds, or whatever: the principle remains) of 1s with
no 0s or vice versa, there are no transitions, and the receiving
clock/oscillator can drift "out of sync" - especially if the propagation
medium is one over which the transition time varies due to physical
factors anyway. "Bit stuffing" - in its various implementations -
ensures there is a transition at least every x bit periods, to keep the
receiving end in sync..

DC removal/AC coupling: There are many physical parts of the system
where the signal _must_ have transitions to work: it may be better to
think of the bits as being + or - rather than 1 or 0. In order to detect
which is being received, the receiving end must maintain a threshold at
the half level - so that it can tell whether an incoming bit is above or
below that threshold, or one side or the other. Because physical systems
drift, it's a lot easier (in the limit, it's only _possible_) if the
signal at the receiving end is continuously switching between the two;
if a long run of one bit value is received, the threshold may drift, so
that either the tail end of the long run is interpreted as the other
bits, or the next real transition is missed. (There are also technical
reasons why it's _easier_, even if not actually _necessary_, to "AC
couple" stages.)

--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Reality television. It's eroding the ability of good scripted television to
survive. - Patrick Duffy in Radio Times 2-8 February 2013