4 Sep
2006
4 Sep
'06
9:49 a.m.
Christian Corti <cc at corti-net.de> skrev:
On Sun, 3 Sep 2006, Johnny Billquist wrote:
try it
> > No, you'd just have to write it from
scratch. Nothing strange
about that, and
> > doing something about V4 and V5 games
isn't that difficult
either. Given a
> > little time I sure could whip one together,
but for now I'll
leave the
> > exercise to someone else.
> > I've already written one Z-machine interpreter in MACRO-11. It
deals
with
> > anything V1 to V8, except for obvious
limitations (no sounds, no
graphics, no
> mouse...)
There I completely disagree! I've ported the Z-machine interpreter from
InterTaskForce (written in C, and really good if you want to understand
how the interpreter works) to the IBM 5110 (written in PALM assembler),
including paging etc. But until now nobody else has ever wanted to on his 5110/5120... Seems all 5110 owners don't
want to use their
machines.
:-)
However, have you ever programmed on a PDP-8?
If so, you would realize that it really requires you to do things a bit
different from what people nowadays do.
You have 4K word pages. One word is 12 bits. The addressing style of
instructions really limit most routines to what you can fit into 128
words. No stack...
Well, there are a lot of things to play around with.
Like I said, it wouldn't be difficult to write a Z-machine interpreter
for the PDP-8, but you'd better not base it on something written in C.
You might pick an idea or two from Frotz, for instance, but you'd have
to rewrite that as well to fit with the PDP-8.
And you have almost no OS to support you either, so you'd have to
implement the I/O as well, and figure out which, if any, clock you have,
if you want to implement timed input.
Johnny
4 Sep
4 Sep
2:09 p.m.
Johnny Billquist wrote:
However, have you ever programmed on a PDP-8?
If so, you would realize that it really requires you to do things a bit
different from what people nowadays do.
You have 4K word pages. One word is 12 bits. The addressing style of
instructions really limit most routines to what you can fit into 128
words. No stack...
Well 12 bits holds 2 6 bit chars nicely :)
I suspect the best way to do this is to convert Z-machine
data into something better suited for 12/24 bit processing.
Well, there are a lot of things to play around with.
Like I said, it wouldn't be difficult to write a Z-machine interpreter
for the PDP-8, but you'd better not base it on something written in C.
You might pick an idea or two from Frotz, for instance, but you'd have
to rewrite that as well to fit with the PDP-8.
Well the PDP-8 is easy to program in -- 8 instructions types. :)
And you have almost no OS to support you either, so
you'd have to
implement the I/O as well, and figure out which, if any, clock you have,
if you want to implement timed input.
Oddly you have better OS support now with the emulators and
some new-ish hardware that uses ide drives.
Johnny
5 Sep
5 Sep
2:54 a.m.
On Mon, 4 Sep 2006, Johnny Billquist wrote:
However, have you ever programmed on a PDP-8?
Not much, but I'm a bit familiar with it.
If so, you would realize that it really requires you
to do things a bit
different from what people nowadays do.
You have 4K word pages. One word is 12 bits. The addressing style of
instructions really limit most routines to what you can fit into 128 words.
No stack...
What's a stack? ;-))
(Note: There's no stack on the PALM either, but other goodies instead, I
admit).
Like I said, it wouldn't be difficult to write a
Z-machine interpreter for
the PDP-8, but you'd better not base it on something written in C.
You might pick an idea or two from Frotz, for instance, but you'd have to
rewrite that as well to fit with the PDP-8.
I think if you take an assembler program as a starting point you might
oversee the essential point of the routine because you are too busy
figuring out how to port that to another assembler language for a
completely different architecture. A (simple!) C program shows you what
you need to know and then you can think of doing the same in whatever
language and for whatever architecture. That's the reason I chose the ITF
interpreter (and not Frotz or any other one).
And you have almost no OS to support you either, so
you'd have to implement
the I/O as well, and figure out which, if any, clock you have, if you want to
implement timed input.
Well, there's no OS at all on the IBM 5110, only some monitor routines
that assist you in accessing the floppy or the printer, but the I/O (and
much else) has to be done from scratch (there are no library routines,
nothing!). You have to write your own 16 bit math routines, multiplication
and division and the like. Although the PALM is a 16 bit processor, the
ALU only is 8 bits wide (with carry into the high order byte). There are
neither screen output nor very handy keyboard input routines, so you need
to get a bit creative. Ah, and not to forget the Z-Code <--> ASCII <-->
EBCDIC conversions required, custom line wrapping (screen has only 64x16),
custom "MORE"-like pager, etc.
Christian
6466
days inactive
6467
days old
2 comments
3 participants
participants (3)
-
bfranchuk＠jetnet.ab.ca -
bqt＠softjar.se -
cc＠corti-net.de