Re: Sockets, port and loop-back ?

From: William D. Tallman (wtallman_at_olypen.com)
Date: 01/09/04 

Date: 8 Jan 2004 17:58:45 -0800

ptb@oboe.it.uc3m.es (P.T. Breuer) wrote in message news:<95cftb.um5.ln@news.it.uc3m.es>...
> William D. Tallman <wtallman@olypen.com> wrote:
> > Well, fancy that: the very next thread keeps the subject alive. What are
> > the odds? <grin>
>
> Hmmm.
>
> > Dammit, I'm not seeing Peter's posts!! I'm gonna have to get another news
> > server!!! Argh! This is ridiculous! And I'll bet Peter can't see this
> > response, either... <sigh>
>
> No, he can see it. But he hasn't seen much of his own posts in this thread
> either. Maybe he's killfiled himself for rudeness? He's a
> schizophrenic, you know!

Peter, you are your own worst critic, you know. Happens to all the
best people, though.
 
> > > well, someone wrote:
>
> > That was me!
>
> Now I'm confused. I thought that was you.

Not you, me! Who's on third?
 
> > Okay, one has to remember that for *nix operating systems, *everything* is a
> > bit-stream that is identified as a file. *nix knows nothing about
> > hardware.
>
> Bit like some people, really.

Yeah, but they know about hard bodies, I bet.
 
> >Therefore, networks are like devices, they are conceptual
>
> Arrgh ... metaphysics alert! Danger! Danger!

Careful!! The Great Penguin might hear you!
 
> > entities only. After all, an operating system itself is only a conceptual
> > entity, the manifestation of which is a collection of digital information
> > chunks stored here there and yon in a computer.
>
> Yerrrssssssssssss. So it's all a fake, at bottom? Like the moon
> landings? Or W Bush's plan for the reconstruction of Iraq?

I think that may be so. Took a look inside a dead drive once, and
couldn't see a thing other than dead drive parts. Also took a "dead
bug" apart once. Smashed it with a hammer, actually. Was left with
"dead bug dust", that's all.

But I never was sure it wasn't lunar dust, though. Could have been
ziggurat dust, as I was using one of those lunatic newspapers about
ancient astronauts who actually built the things.

Ummm... about "The Shrub". He is exactly what the population of this
country decided they needed, and thus exactly what they deserve.
Perhaps you can imagine, though, the anguish of a few U.S. citizens
that they should be represented by this pin-headed twit. Great thing
about American politics: such as he can only last so long. Terrible
thing about American politics: the populace regularly delivers itself
of its pent-up stupidity right about election time, and such as The
Shrub manage to get in office.

But that's getting dangerously close to reality.....
 
>
> > When the hardware starts
> > access to those chunks, it finds instructions about how to do things, but
> > it's the hardware that does the doing, not the software.
>
> Now I'm confused again. I thought it was the software that pretended
> tobe hardware? So what is the hardware? Really a conceptual entity in
> the mind of the software?

That also may be the case. Probably a case of "hard**** envy", you
know. One of those kind of dreams...
 
> Are you sure you didn't use to be a scriptwriter for Star Trek? Or
> don't you use matchboxes (for writing script plots on the side of)?

Oh no!! Heaven forbid!! Was considered for Dr. Who though. Loved
all the tin foil and cardboard props....
 
> > So, take a look at the OSI model, and find that the bottom layer is the
> > physical reality.
>
> And I thought it was buttered toast. The buttered side always lands on the
> bottom when I test.
 
Nah, that's electron grease you're seeing there. Makes for faster
networks.
 
> > We are so used to assuming that anything that has
> > existence
>
> Hey, non-existence exists, doesn't it?

Well, now if you consider the "Many Worlds Hypotheses", it probably
does somewhere.
 
> >also has physical manifestation that we have difficulty in
> > contemplating the matter of networks apart from the machines that use them.
> > We assume that for any network, there must be a physical device that uses
>
> Is this like the tree in the quad?

Which quad? We have several over here. The tree in the Stanford quad
was called "The Palo Alto", but they transplanted it elsewhere and
named a town after it. What others?
 
> > it. The result is the confusion of a network address and a NIC. The NIC
> > doesn't have a network address, though it's firmware hosts a MAC (machine
> > access code). Which is why ARP (address resolution protocol) exists, and
>
> Please don't burp.

That's not a burp. It's a Hahvahd "urp", which is a precursor to
digital projectile... well, never mind.
 
> > why a tcp/ip dump is fraught with "who has" requests.
>
> > The IP protocols exist a couple of layers up the model, and have nothing to
> > do with how links are established, or how how electrons are continually
> > inconvenienced at our behest.
>
> Phew! That disk with a TCP chip on it that I saw the other day WAS a
> hallucination, then!
>

No, don't think so, particularly if it was wearing it on its shoulder.

> > The IP is a routing protocol, which deals
> > with addresses and such. It can be used anywhere that its function is
> > appropriate, and one of those places is where an operating system needs to
> > have an orderly means of interacting with applications. In this regard,
> > there are no physical devices other than the computer's bit-bashing-ware.
>
> Not even NICs? Perhaps NNICs? Not Network Interface Cards? Or NINCs?
> Network Interface Non-Cards?

Careful how you decline NIC.. It is:

                Cingular Puerile

eth0 NIC NNIC

eth1 NICs NNICs

eth2 NINCs NNINCs

Remember to adjust for drift in declination, though.
 
>
> > And this network is called a loopback "device", which is not a physical
> > device at all, but a conceptual communication system.
>
> Ahaaaaaaah. You mean it talks to itself, and calls that a network.

Yep. Solipsisticism abstracted.
 
> > Now, what happens when one wishes to connect the loopback to another
> > network?
>
> A Jolly Good Question. Maybe one needs a Router? :-).

Why? With all the bridges that abound, Cisco never built a router
that could stand in the same bit-stream as the Golden Gate Bridge.
>
> > In the C program language, they have a term called Undefined
> > Behavior.
>
> Really? I just tried compiling a the function Undefined Behaviour, and
> it refused to compile! Or at least I think it did. Maybe it fractured
> the Universe and in this timeline I believe it didn't compile, while it
> really did?

Well, the ultimate authority there is probably your friendly local
rhinodaemonologist. Dunno what they call them in other universes,
though.

>
> > Which means exactly what it says, behavior that is undefined:
>
> Oh, good! I'm glad you defined that!

To define.... the undefinable..... to compile... the
uncompilable....

Sounds like a good song to me....
 
> > it can be anything including exactly what one might expect.
>
> But if it follows your definition, then it is defined, no? And
> therefore not what you said it was!

Can't tell yet. Haven't got the second verse to the song.
 
> > A legendary
> > code boffo put it succinctly by asserting that one indulgence in such
> > behavior caused demons to fly out of his nose!
>
> Oh - that was in Universe #100026678200ab.

Yep, that's the one. Discovered in an unlogged flight of the
DeathStar 9000.
 
> > The specialty that
> > subsequently arose is technically known as Rhinodaemonology.
>
> Well, miracle science also has something to say on the subject. A
> miracle is the programming construct that has "true" as precondition
> and "false" as postcondition. So there is absolutely no difficulty
> in setting up for the call to a miracle, and after it has run,
> absolutely anything at all is true (even "false"), which is very
> convenient.

But the problem here is that you can never change the condition,
unless you use a pointer to the condition. It's not allowed to use
the term 'pcondition', because that can be confused as to whether its
p(re)condition or p(ost)condition. And the term conditionptr is just
so declasse, dontchaknow..
 
> > Perhaps we
> > might consult these good people (rhinodaemonologists) to discover if they
> > have any idea what might happen if one managed to connect a loopback
> > network to a LAN, or some such!
>
> I would sneeze.

Well, that's what they would want you to do. How else to detect the
presence of sulfur?
>
> > In short, serial ports and network ports are, at the very least, not the
> > same thing.
>
> Oh, is THAT the message!! Ha ha ha. Yes, very good. Hilarious.
> Excellent. Superb.
>
> Remember to rule out ports with boats in too. And ports with air in.

Well, not so fast there! Ports are bidirectional by default, you
know, and the boats out and air out live in the other half of the
duplex, and they're problematic. Heisenberg said you couldn't talk
about boats (or air) in and out at the same time. Seems there's a
Planck connecting the two halves of the duplex....
 
> > A serial port is a physical connection device that spits out a
>
> It doesn't sneeze?

Oh no, that's so unsanitary! That allows bugs in the system, and one
never knows if one's debugger is up-to-date enough to handle such
things.
>
> > bit stream (also receives same). A network port is the final part of an
> > IP4 address.
>
> Hmmmmmm.

[you're right. there's a missing 'v' between 'P' and '4']
 
> > Well, that's off the top of my head. I've lost some hair up there, and so
> > some parts of this blather may also be missing. Anyone discovering such
> > errata is invited to report same in response!
>
> I don't think I have the heart. Whatever you are taking, it is at least
> 15 years old, and smooth!
>
> Peter

Now Peter, you know my tastes better than that. It's from the Cask in
Poe's basement. At least a century old... well, that *was* last
millennium, was it not?

Nice to hear from you again!

ROFL!!!!!!!!!!!!!!

Bill Tallman