Re: Schneier, Data and Goliath: no hope for privacy

On Sun, 23 Feb 2025, Salvador Mirzo wrote:

It saves all messages in a local spool folder, and since nntp is a
nice and simple retro-protocols, it is trivial to understand the
format. So what you could do, between 2 leafnode servers, is to just
reverse engineer the format and "copy" the spool directory between the
two leafnode installations and all the messages will pop up on the
other leafnode as well.

>
Okay, but the question was to just to confirm my mostly-forgotten
recollections of Leafnode.  I wouldn't mind working on it to make it
peer via NNTP itself.  But I would much rather write a completely new in
a non-C language.

I wonder if there are any good C to Go converters out there? Would be
interesting to see how much effort it would take to convert leafnode from c to
go? Maybe then, it would be an easier code base to work with?

I would be surprised if INN2 doesn't do all of this, but I think we
should have other alternatives with newer ideas too.  For instance, I

>
The reason I did not go the INN2 route was that I wanted some small
and simple, for pulling messages one way only. Leafnode at first had
some restrictions such as needing a valid DNS name which was a pain,
so I simply deleted its checks, and now the server can be named
anything I want, which is fine, since I'm not peering. This can of
course be added back if you want it.

>
I never used INN2, but I do suspect that it's made for a serious USENET
server and that it's more complex that it needs to be for the idea of a
network of small NNTP servers.

Yes, I think it is way more complex than leafnode.

I think is perhaps somewhat of a downward trend. I feel awe when
talking to the older generations who had to learn the hardware,
program in assembler and so on.

>
I feel the same.  Like you, I feel great learning from the older
generations.  In fact, I often think that they were privileged for being
able to be there first.  I identified this easily enough to develop a
passion for studying the history of computer science, which makes me
look very old now because I use a lot of very old tools, which are
awesome tools despite their age.  I got a web post by Joel Spolsky the
phrase that ``software doesn't get dusty''.

True. I have a retro-class on thursday and will show them some nice stuff in the
form of vim, alpine, and midnight commander. Apart from a shell (bash) those are
my main tools in the terminal.

In my generation, hardware and assembler were solved problems, so the
programming was done in higher level languages.
>
Todays generation don't even see the hardware, they all use cloud
servers and python.
>
So the original foundation gets further and further away. Only a small
set of hw wizards still care and know about that layer of the stack.

>
That's quite right.  I went through the same.  The whole thing was
pretty much already done.  I believe I am not very fond of directly
interacting with hardware myself.  For example, I usually like to have a
very clean office---no wires (if I could), not a lot of gadgets around.

When I was young, as a system administrator, I loved wires all around, disk
drives, NIC:s, power supplies etc. My office would look like a junk yard. But as
I grew older and moved more and more into sales, I now have younger guys who
have taken over that role, and that messiness. ;) My office now has one cable,
and that is the power cable. Since I get about 13-14 hours out of my laptop, I
don't even need a power cable during the working day, unless I have many hours
of video calls on that day.

Nevertheless, I feel obsessed by computers and I try to get close to the
hardware by more abstract means.  For instance, I've been reading about
the 6502 and it seems like such a simple CPU that it makes up for a very
great computer architecture first introduction, unlike x86, say.

I remember programming for the Z80 when I was young, on my calculator, and also,
of course, assembler on the 486. Those were the days! =)