Re: Problems with maXZIP 1.3..
11 Nov 1995 09:52:57 +0100
In article <47v7h0$g05@news.ox.ac.uk>,
David Kinder <kinder@teaching.physics.ox.ac.uk> wrote:
>Andrew C. Plotkin (erkyrath+@CMU.EDU) wrote:
>: Also fix the problems with Z-machine interpretation; the
>: time delay in timed input is ten times too long. (This is a bug
>: inherited from ZIP.)
>
>Perhaps now the bug is in there it's best left. Otherwise every game that used the
>timed single key-press opcode would have to come in two flavours; one for the fixed
>interpreters, and one for those that haven't been fixed. Besides, do we really need
>1/10th second timer resolution? Perhaps it should only be fixed for the V8 spec?
There are only two games using timed input (as far as I know): Border Zone
and Freefall. Border Zone expects 1/10 s but also runs on the old ZIP
(for reasons explained in the new Specification), and Freefall runs just
fast enough under 1/10 s. Future authors of games with timed input will
surely read the Specs before they program this feature, so they will
know that 1/10 s is correct. All the original Infocom interpreters
implement the feature correctly; only ITF and ZIP have the bug and these
can be fixed easily on all platforms. So I don't see any reason to stick
with the old (incorrect) implementation.
-- Dave