Simple but fast control

Geoffrey Clements

"Paul Martin" <pm@zetnet.net> wrote in message
news:slrnfgkijd.gvr.pm@thinkpad.nowster.org.uk...
> In article <470a4053$1_1@glkas0286.greenlnk.net>,
> Geoffrey Clements wrote:
>> "Daniel James" <wastebasket@nospam.aaisp.org> wrote in message
>> news:VA.000011d0.437d5233@nospam.aaisp.org...
>
>> [snip]
>
>>> Windows has always provided some form of multitasking capability, and a
>>> clipboard to cut/paste data between applications. I remember doing that
>>> on DOS by running one app after another with SideKick in the
>>> background, and using SideKick's screen-grabbing capability to copy
>>> data between the apps. Windows's clipboard certainly makes that much
>>> easier.
>>>
>
>> Ugh! - you've just reminded me of TSR (Terminate and Stay Resident)
>> programmes. I remember trying to write a couple of these in assembler
>> and
>> my palms are starting to get sweaty - scary stuff!
>
> Trying to write!
>

Yeh - it was just a bit of fun, I have this vague memory of trying to slip a
new interrupt vector into a chain of interrupt vectors without having all
the information I needed. Sometimes the PC would crash immediately and
sometimes it took a few seconds :-)

> Our business depended on a TSR (written in Turbo Pascal) sitting under
> Desqview, acting as an IP multiplexer on top of the Crynwr packet
> drivers. It even did ARP. That was in the days when we ran physical
> modems, four to a motherboard.
>

hehe - glad I didn't have to support that! I did play around with Turbo
Pascal though, ISTR it had a remote debugging facility which was
particularly nice but I may be confusing it with something else. More
memories surfacing - OWL (Object Windows Library?)

--
Geoff

Folderol

On Tue, 25 Sep 2007 15:58:29 +0100
Tony van der Hoff <news@nospam.vanderhoff.org> wrote:

<snip>

> Hmm, they don't seem to be too keen on people finding the licence conditions
> on their web site; where did you find it?

http://www.microware.com/corporate/copyright.cfm

it's not clear *exactly* what that applies to

--
Will J G

Folderol

On Tue, 25 Sep 2007 14:04:53 +0100
Martin Gregorie <martin@see.sig.for.address> wrote:

> Folderol wrote:

<snip>

> > Thanks for the suggestion. I had a quick look, but it seems
> > over-complicated.
> >
> Its no more difficult to program than writing C for Linux. The IPC stuff
> is just different, not harder.

As I said, I only had a quick look. I'm currently looking at several
different options. One thing I especially want to mitigate if at all
possible is hardware obsolescence. So I want to try and build
flexibility into the design.

> The OS itself is a lot simpler than Linux and has much lower hardware
> requirements. My system has a whole 8MB of RAM and has never come near
> to filling that up despite its 1.4 MB RAM disk.

One problem is that it seems I need a windows box to run the software
that runs the software if you get my drift. Or have a missed something
here?

--
Will J G

Paul Martin

In article <470a4d58$1_1@glkas0286.greenlnk.net>,
Geoffrey Clements wrote:

> "Paul Martin" <pm@zetnet.net> wrote in message
> news:slrnfgkijd.gvr.pm@thinkpad.nowster.org.uk...

>> Our business depended on a TSR (written in Turbo Pascal) sitting under
>> Desqview, acting as an IP multiplexer on top of the Crynwr packet
>> drivers. It even did ARP. That was in the days when we ran physical
>> modems, four to a motherboard.

> hehe - glad I didn't have to support that! I did play around with Turbo
> Pascal though, ISTR it had a remote debugging facility which was
> particularly nice but I may be confusing it with something else. More
> memories surfacing - OWL (Object Windows Library?)

That's Borland Pascal -- much later.

--
Paul Martin <pm@zetnet.net>

Daniel James

In article news:<slrnfgk5c4.eo7.pm@thinkpad.nowster.org.uk>, Paul
Martin wrote:
> Hence the demise of OS/2 -- far, far better designed than Windows,
> hampered by the need to run it as a virtual machine, and killed by
> Windows 95 coming on the scene.

OS/2 was killed by enormous ineptitude on the part of IBM (both
technical and marketing) and by the fact that NT beat for kinds of crap
out of it before breakfast.

At one time I was working with NT 3.1 Beta and OS/2 2.0 Gold running on
two computers side by side, and the box running the Beta software was
NOT the one that was always falling over and needing to be rebooted.

OS/2 was far better designed than either Windows 3.x or Windows 9x,
with their underlying dependence on creaky old DOS, but NT is a
different cauldron of lobsters altogether.

Cheers,
Daniel.

Daniel James

In article news:<470a4053$1_1@glkas0286.greenlnk.net>, Geoffrey Clements
wrote:
> Ugh! - you've just reminded me of TSR (Terminate and Stay Resident)
> programmes. I remember trying to write a couple of these in assembler and
> my palms are starting to get sweaty - scary stuff!

Pah! Nothing. I wrote a DOS device driver that could load a TSR program from
config.sys.

[You tell that to young people these days ... ]

There was a reason. A client of mine had a device driver that carried on a
dialog with the user as it loaded, and one of their customers wanted to be
able to use Portuguese language and characters in that dialog. the problem
was that the Portuguese codepage was normally loaded (as a TSR) from
autoexec.bat, and so was not loaded until after the drivers mentioned in
config.sys had all already loaded. So, I wrote a device driver that could
load and run the MODE application to load codepage 860 from config.sys
/before/ my client's device driver loaded.

You could ask it load just about any DOS app ... but I didn't provide more in
the way of memory management than was needed by a simple one-segment
application that would either terminate normally or terminate and stay
resident ... so running multi-segement .exe files, anything with overlays, or
anything that spawned another program was pretty-much a non-no.

[Ah, them was the days ... pass me another bottle of Chateau de Chastelet,
will you?]

Cheers
Daniel.

Daniel James

In article news:<9dvpt4-6cl.ln1@zoogz.gregorie.org>, Martin Gregorie
wrote:
> Yep. OS-9 actually precedes DOS as well as being considerably better -

Yes, I recall that I was running CP/M-80 on a Z80 at around the time I
first heard of OS-9. I was starting to thing that the 6809 looked like a
nice chip with a nice instruction set and perhaps that was the way I
should go ... when the world suddenly started going 16-bit around me.

> ... DOS didn't even have directories before v3.

V2.0, actually, but who's counting?

> Unless you're using a kernel with swapping removed you can never
> guarantee deterministic response times (a common problem with all
> virtual memory OSes including 'doze if you're writing time-critical
> real-time systems).

True ... but linux does give you the freedom to customize the kernel if
you need to (at no extra cost -- in money, at least). I assume that anyone
who does have an unusual requirement of that sort isn't going to pale at
the thought of doing what's needed to satisfy it (an assumption that may,
I agree, prove unwarranted).

> In addition AFAICR the standard Linux/UNIX task scheduler uses a simple
> priority/preemption algorithm while OS-9 has a more sophisticated and
> more tunable time sliced algorithm of which priority/preemption forms a
> subset.

Again, the standard linux kernel sources do allow for some customization
(though I'm not sure of the details, as I've never needed to fiddle with
this aspect of kernel configuration). It's also possible to replace the
scheduler entirely, of course (as was pretty-much done AIUI between the
2.4 and 2.6 kernels) but I'm sure that is not for the faint-hearted.

Cheers,
Daniel.

Tony Mountifield

In article <03ivs4-jvs.ln1@zoogz.gregorie.org>,
Martin Gregorie <martin@see.sig.for.address> wrote:
>
> Radisys European support is in Germany if you want to talk to them. Tony
> Mountifield may know of some UK support: since Windrush vanished I no
> longer have local commercial contacts.

I'm afraid I don't know the current OS-9 world at all. I looked on
Radisys' site the other day and saw that Wordsworth Technology are
listed as a reseller in the UK, but I don't know whether that is just on
their own hardware. They were an OS-9 customer when I was at Microware -
I wonder if Kevan Wells is still in charge?

Cheers
Tony
--
Tony Mountifield
Work: tony@softins.co.uk - http://www.softins.co.uk
Play: tony@mountifield.org - http://tony.mountifield.org

Daniel James

In article news:<03ivs4-jvs.ln1@zoogz.gregorie.org>, Martin Gregorie
wrote:
> OS-9 is a full OS in its own right and runs stand-alone on PC
> hardware. It just won;t need much of it. I'd expect it to run on
> the low power mini-IPX boards for example.

Not that that's saying much ... even the lowest-powered mini-ITX boards
will run Windows or linux (though they don't really hold enough RAM for
Vasti). OS-9 should run quite nicely on much more modest hardware than
that.

I remember thinking how nice OS-9 looked back when it ran on 6809s, but
I never actually invested in any hardware to run it on. Nice to see it's
still around.

Linux is cheaper, though ...

Cheers,
Daniel.

Paul Martin

In article <VA.000011d6.48b9f4f5@nospam.aaisp.org>,
Daniel James wrote:
> In article news:<slrnfgk5c4.eo7.pm@thinkpad.nowster.org.uk>, Paul
> Martin wrote:
>> Hence the demise of OS/2 -- far, far better designed than Windows,
>> hampered by the need to run it as a virtual machine, and killed by
>> Windows 95 coming on the scene.

> OS/2 was killed by enormous ineptitude on the part of IBM (both
> technical and marketing) and by the fact that NT beat for kinds of crap
> out of it before breakfast.

> At one time I was working with NT 3.1 Beta and OS/2 2.0 Gold running on
> two computers side by side, and the box running the Beta software was
> NOT the one that was always falling over and needing to be rebooted.

Don't forget that OS/2's HPFS was contributed by Microsoft, back when
Microsoft and IBM were best buddies.

--
Paul Martin <pm@zetnet.net>