Re: Nobody proposing new uk newsgroups - why ?

On Sat, 17 Dec 2005 16:21:41 +0000, Dave J
<requiem@xxxxxxxxxx> wrote:

> In MsgID<X6c9yeWjHyoDFw39@xxxxxxxxxxxxxx> within uk.net.news.config,
> 'Molly Mockford' wrote:
>
>>At 15:54:04 on Fri, 16 Dec 2005, Dave J. <requiem@xxxxxxxxxx> wrote in
>><gbm5q1d3sprapn69akb21onobr1s1hg13n@xxxxxxx>:
>>
>>>In fact, it took longer to download (on dialup) than it did to install and
>>>initialise. And that was as a complete newbie who (almost) still pictured
>>>TCP 'ports' as seperate pieces of wire coming out the back of a machine.
>>
>>You mean they're not??? (There's an awful lot of pieces of wire coming
>>out of the back of my machines...)
>
> Reason I made that comment is that on my first venture onto the internet
> (via the university computer department long after I'd ceased to be a
> student) that genuinely was how I was picturing it. It confused me that so
> many ports could be open on one machine.

Nothing wrong with it as a picture.

> I pictured TCP machines as having like a 'filter box' (maybe I thought
> that's what a router was) that separated out the different port requests
> and fed them down separate pieces of wire to the computer. Don't forget,
> my *only* idea of a port was a physical serial or parallel port..

Yup. And as far as users and most network application programmers are
concerned they might as well be exactly that. I use that mental
'picture' about ports, and pipes, and other communications channels.
Most people, including programmers, don't need to know anything about
the actual protocols, a 'picture' of it as actual holes (in a firewall,
for instance -- I hear admins talking about "poking a hole to let
$application through"), pipes and bits of wire is as good as any.

> Of course, my readings at the uni (telneting to web servers and saving the
> logfile for later translation as I didn't have access to their web
> browser) disabused me of that notion fairly quickly, but I do remember it
> with great amusement.

I remember confidently predicting that it would be impossible to get a
computer on a chip because of all the wires which would be needed. I
hadn't thought of multiplexing. Of course, with the pin/contact density
of modern processors (900+ on an inch square!) one could almost do it
without multiplexing for something the complexity of a 1960s computer...

Chris C
.

Relevant Pages

  • Re: var keyword in SystemVerilog
    ... wire S my_wire_s; ... data type ON ITS OWN mean you get a variable, ... var S my_variable_s; ... input ports are always nets: ...
    (comp.lang.verilog)
  • Re: 1-wire to USB converter that can use 1820s directly
    ... than to need to add another PCI card to support serial ports. ... how someone could finally say "not one more f'ing wire!" ...
    (comp.home.automation)
  • Re: Sony KV J29 Green Lines
    ... to 45 sec for the picture to appear on the screen. ... Trace out the green video amp, try disconnecting the wire carrying the ... The sony technician said he would do ... One wire end is soldered to the H pin on ...
    (sci.electronics.repair)
  • Re: Nobody proposing new uk newsgroups - why ?
    ... >>> I pictured TCP machines as having like a 'filter box' (maybe I thought ... >>'picture' about ports, and pipes, and other communications channels. ... pipes and bits of wire is as good as any. ... you can picture the 'inside of the box' in any way you ...
    (uk.net.news.config)
  • Re: ICOM AT-150 tuner antenna ports
    ... where the impedance mismatch between the wire and the rig/tuner isn't ... long-wire antenna port on a ham-band tuner. ... mount and wire in another S0-239 connector and have 4 working ... when all the other ports ...
    (rec.radio.amateur.equipment)