12 Sep
2007
12 Sep
'07
11:18 a.m.
On Tuesday 11 September 2007 07:50, Allison wrote:
Yup...
I remember those days, thinking of 50 nodes as being pretty good-sized. :-)
What does that take on the software side of things?
Not a whole lot, CPnet could be used but it was easy enough to use plain
vanilla CP/M2.2 and add your own BIOS drivers for "networked functions".
defineatly home grown.
Modifying the BIOS of a machine was something I'd never quite gotten as much
of a handle on as I'd like to. I still have a Bigboard II that I didn't get
the software with, for example. While it'll boot a Xerox 820 floppy, that
only gives me SSSD, the DS or DD stuff won't work because the two versions
of things map the I/O differently (the BBII from 80H rather than 00H). I
need to get back to hacking on that one of these days.
Just that I'd like to see some stuff that isn't oriented that way. You
have a SBC, you obviously need some way to talk to it, but the standard
"console" stuff gets a little old, I probably don't want to hook a
printer up to it, and may not even want a disk drive of any sort,
depending on what I wanna do with it. I'm up for exploring some
alternative approaches to doing things. Unfortunately the embedded stuff
that's out there doesn't satisfy too often, the design being too specific
to the app, source code not available, etc. I'm thinking that it should
be possible to have some sort of a more generalized framework to hang
things on, and then you could optimize it for specific uses, or expand
it in different directions. Even from the earliest days "personal"
computers all seemed to take pretty much the same approach to things...
Well by hook or by grook the average PC still has a serial port, some have
two or atleast a USB port for a USB to serial.
Actually amongst my stuff USB is a really recent phenomenon, and I've only in
the past month or two acquired anything that uses it -- an external DVD
reader / CD burner -- and nothing else. Most of my peecee hardware is older
stuff, and serial ports are not a problem, by comparison with those whose
hardware is more current. So yeah, having a serial port is probably one of
the things I'd consider to be a necessary feature in there. And it'll end up
being able to talk to just about anything I'd be running.
On Sunday 09
September 2007 16:04, Allison wrote:
I remember seeing that in some sales literature and it always did strike
me as being more spendy than I wanted or could afford to get into. :-)
It cost more because it was industrial strength, larger boards, regulated
power and so on. A while
back I *almost* got a hold of one of those "z80 network in a
box" systems, it wasn't S-100 but something else I can't recall, I
think that's the one I have the book on, but I never did snag it.
Multibus, very nice bus and expensive cards. I have a few multibus
cards. Intel used it in their MDS800 and a few otehrs as well. I know of
ARCnet, went to a short seminar on that once at a trade show,
and in fact even have a couple of ISA cards around here someplace,
though I don't forsee me ever using them.
ARCnet and most of the 'nets were in the price range of a hard disk then.
Also the whole idea of networking was new. For example in 1982 the two
largest networks I knew of were DEC (internal) and Dupont(internal) and
they were around 50 nodes! There were a few simple schemes but excluding myself
how many hobbiests
back then had two or more systems?
Good point. It took me quite a while to progress beyond my first machine,
and even then it was CP/M boxes that people didn't want any more, rather
than newer/bigger stuff. And I still have most of 'em. :-)
I've also
seen some "CP/M networking" stuff referred to that was
supposed to work through serial ports, which pretty many machines
had, althogh they appeared in at least one case to be using diodes to
wire-OR RS232 signals, which doesn't strike me as too terribly robust.
And what software support there was for this wasn't real apparent.
That was a poor mans networking. Basically the serial ports were used
as CD/CSMA bus and there was some protocal like Ethernet but slower and
could use the usually common async chips. I have such a net going for
my CP/M crates and all. TurboDOS is
neat, and has some good design aspects in it, but there's
too much legacy stuff in there for being able to run CP/M software,
stuff I'd leave out if it were me and too much emphasis on the same
old Console / Printer / Disk Drives in the system, as opposed to
something different or unique. I found the same thing to be the case
when I looked at FORTH, too much of the usual stuff, and that was
supposed to have been used in some control applications? I must've
missed something there...
???? Whats the question or point? It's not that hard to write software to use that
serial as a access from the
SBC for things like a remote printer or disk and people have and are doing
it. It's not "networking" in the full blown sense but none of the IO of a
SBC is required to direct connection to a printer or terminal (ignoring rom
based stuff).
RF is a whole 'nother thing entirely, and I've not even begun to do much with
that...
Yet. :-)
--
Member of the toughest, meanest, deadliest, most unrelenting -- and
ablest -- form of life in this section of space, ?a critter that can
be killed but can't be tamed. ?--Robert A. Heinlein, "The Puppet Masters"
-
Information is more dangerous than cannon to a society ruled by lies. --James
M Dakin
I realized in other messaging a while back that
it's been well over a year
since I fired up a soldering iron, and this is a bad thing. :-) And
even then, it was a matter of scrapping stuff, not building anything
new and interesting. I need to get out of that particular rut and get
back to it, or there's no point to all those parts I've been scrounging
for decades. Maybe one of these days I will...
When you do tell us about it. Seems these days I get to maybe one
of the major computer construction based projects maybe two per year.
But I split my time between RF projects and digital projects.