> > Yes, but it some cases the 'kludges' led to problems later on - 'trivial'
> > problems like a total incompatability between the Disk II and just about
> > any other machine in the world
>
> For the most part, 5.25" disks are incompatible between any
> two machines. There never was a standard format, unlike the 8"
> world where IBM 3740 dominated (though that didn't mean that
> everyone was compatible with it.)
Yes, but in general you could get 2 machines that used FM (or MFM)
recording on 5.25" disks to agree on a common format to interchange data.
I got my TRS-80 Model 1 and RML380Z (a UK CP/M machine) to transfer text
files that way.
>
> > a PSU that was beyond the design limit when running a system
> > board, language card and 1 drive
>
> No, it isn't. I just measured the system current drain (II+, 1 Disk ][
> drive, third-party language-card equivalent) and verified
> this isn't true. It is true that every 10 years or so a power supply
> will fail, but you can blame Astec for that.
Maybe my motherboard is marginal, but I remember measuring the +5V drain
of motherboard + Apple Language Card + disk II controller + Disk II as
(I think - this was about 5 years ago) 2.6A. According to the label on my
PSU, the maximum 5V drain is 2.5A. That's marginal.
I have never had an Apple PSU fail, thankfully...
>
> > that _crazy_ slot addressing scheme
>
> That scheme is actually pretty clever. I can put a disk controller
> in slot 5, type "CATALOG,S5" and look at a drive connected to it.
> No DIPswitches to set, no BIOS to modify.
I dislike 'geographical addressing' schemes in general. I'd rather have
all my peripherals at fixed addresses no matter how I put the cards in the
box. Also, geographical addressing limits the maximum number of slots you
can have (something I tend to run out of on all my machines....)
> I can put a printer card in slot 4, type "PR#4", and now all redirected
> to the printer. I can put a serial card in slot 1,
> type "IN#1", and now a keyboard hooked to a terminal on the serial
OS-9, one of my favourite micro OS's has a very nice way of handling this.
Suppose you want to add a serial card based on the 6551. You can have a
_generic_ 6551 driver that has no idea of the address of your card, and
then a device descriptor that contains the address and the paramter table
(baud rate, whether you want a LF after a CR, a pause at the end of a
screen,etc). Although most hackers used the standard assembler to make
descriptors, it would be trivial to make a 'user friendly' installation
tool.
IMHO that's a much nicer way to do it than the Apple method which means
you have to know which slot each card is in (and, in a lot of cases there
are 'standard' slots for particular cards that you'd better stick to. What
do you do if you want _2_ 80 column displays in the same machine? One in
slot 3, what about the other one).
> > and the saving on chips/PSU consumption by switching the power line to
> > the I/O card ROM
>
> Which I/O card is this? I've never seen anything quite like this on
> any of my peripherals (or if I have, I haven't noticed over the past
> 20 years.)
I have certainly seen 3rd party cards that did this. I forget the
manufacturer, but they made a serial card using a real UART and a 9511
arithmetic copro card. I can look it up for you. There was a note in the
manual on how to disable this 'feature' if you wanted to put RAM in the
ROM sockets.
I thought Apple did this on some of their cards. I will have to check my
schematics.
>
> > Saving components is only 'good' when it doesn't affect performance. I am
> > not convinced that this is the case with the Apple ][
>
> Oh - you'd have preferred Wozniak to use single-density FM, getting
> only 90kbytes per 5.25" floppy, over the GCR which lets you get
As it happens, yes I would if it meant I could transfer data to just about
any other machine of the time.
> 140K? Or you'd rather lose some CPU cycles in order to refresh the
> RAM, because your preferred memory-mapped screen layout wouldn't
> allow the video circuitry to do it?
Oh come on. You can scramble the address lines to the _RAM_ so that all
cells get refreshed in a sufficiently short time and still have a linear
address map as seen by the processor if you add a few more TTL chips. I've
seen it done on countless other micros.
>
> Tim. (shoppa_at_triumf.ca)
>
--
-tony
ard12_at_eng.cam.ac.uk
The gates in my computer are AND,OR and NOT, not Bill
Received on Sun Jun 22 1997 - 23:03:52 BST