"Gumshoe" oddities

I wonder if anyone’s noticed. “Gumshoe” must do something quite badly, or quite hacky, with its time management, and I was wondering if there would be an interpreter that would play it properly?

In most interpreters, every action takes two minutes. This is unusual enough that I think it’s a bug - every single action taking two minutes in a game that keeps track of time is unlikely to be by design.

Here’s the cute bit, though: in most interpreters, it displays correctly on the status line as “Time x:xx”. On parchment, though, it’s very amusing: it shows “Score: X Moves: YY”, where X is actually the hour and YY is actually the minutes.

I don’t suppose anyone would know of any interpreter that would make sense of whatever Gumshoe is doing? I recall, for instance, that in a lot of time-sensitive games, ZaxMidlet would infuriatingly make two turns pass instead of just one in various games, which is why I’m assuming it’s an interpreter thing.

Is that a Z3 game? Before Z-Machine version 5 there was a bit in the header which marked whether this was a “score game” (displays “Location: [first global] Score: [second global] Moves: [third global]”) or a “time game” (displays “Location: [first global] Time: [second global]:[third global]”). Perhaps Parchment is not reading this bit correctly, or is treating it as a Z2 game. (The format does allow some flexibility, as e.g. Suspended used the first global for the player character, the second for the number of deaths and the third for the time in minutes.)

I’m all for flexibility. :slight_smile: This is a .z5 game. On the ABOUT banner it shows “Inform v1502 Library 5/12 Standard interpreter 1.2”.

The 5/12 library checks the same header bit, emulating the V3 VM behavior. (An Inform game can set the bit with the “Statusline time” directive – rarely used, but clearly Gumshoe has done this.)

I suppose Parchment is clearing that bit at startup, which I think is incorrect behavior. On the other hand it’s slightly questionable for a V5 game to have it set in the first place.

Incidently, I took a peek at the solutions, which has a table of events, saying at which time the events happen.

I was wrong in assuming the two-minutes-per-turn thing was a bug: strange at it is, the table of events has only even-ended times, so it must be by design.

Either that or the author couldn’t get it to work properly and called it a feature. :wink:

According to the standard flags 1 bit 1 is undefined for versions 4 and 5. It wouldn’t be hard to stop Parchment from clearing the bit… can anyone think of any downsides to making that change?