RT11 / RSTS-E games

[Ethan Dicks]

On Fri, May 8, 2015 at 12:45 AM, John Wilson <wilson at dbit.com> wrote:
> On Fri, May 08, 2015 at 03:47:13AM +0200, Johnny Billquist wrote:
>>Cool. I know that Infocom did publish ZORK I for RT-11, but I haven't seen it
>>myself.
>
> At one point they had RT-11 versions of most of their games.  I bought a
> copy of Sorcerer from them (for $35) a trillion years ago, and disassembled
> and commented it.  I later tried to use it to play Hitchhiker's Guide
> (which was never released for RT-11) but I think it was a microscopically
> different version of the Z-machine so there were problems.

The Z-machine for RT-11 that Infocom released only handled v3 images.
HHGTTG definitely had a v5 version, but without looking, I can't
remember if the first version was v3 or v4.  The difference between v3
and v5 is not microscopic - it's on the order of 4X in terms of object
and string capacity.  ISTR in v4, they added some real-timey "status
line" stuff, for games like "Seastalker", and bumped up some of the
memory addressing limits.

You can run a v3 game comfortably in 48K bytes, painfully in 32K
bytes, even on an 8-bit host.  You essentially can't run v5 game on
32K because the "impure store" and the interpreter won't fit (or it
_might_ fit but with not enough free pages for the parser to remain
resident).  Some of the later games essentially required 64K or even
128K of mem.  Once machines jumped to 256K or more, the paging system
fell away and the interpreters just sucked the entire game file into
RAM and moved pointers around.

As for the PDP-11, one _could_ write a split-I&D interpreter for
larger games (to have ~64K of game/string space) but you'd have to
really want to play a later game on that platform to make the effort
pay off.  There are over a dozen v3 games you _can_ play with 64K of
total space that play just fine.

-ethan