This is mostly to put in a plug for Juhana Leinonen’s new extension “Object Response Tests” (linked from here). The extension implements an ANALYZE verb that systematically runs through standard rules actions (and new actions, if you choose) on any object in your game, exposing infelicitous responses, incorrect plurals and articles, and all those other little flaws that can escape even a good human beta-tester.
I spent a lot of yesterday evening thus: extract a complete list of in-world objects from the world index of Alabaster; write a test script to run through and ANALYZE each of them in turn; revise responses I didn’t like; run the test script again…
Unsurprisingly, this is a lot more thorough than even (what tried to be) a fairly rigorous testing by hand. It’s also a lot more fun. (Though I find it has the perhaps-unhealthy effect of making me lavish a lot more time on commands no one is ever going to try, like DRINK UNDERGROWTH or STAND ON CHIN or TIE FACE TO SKY. There’s something about seeing the same library messages over and over that makes me want to spice them up with randomized elements.)
More seriously: the huntsman seems to be getting a bit more character as I go along. I picture him as a loner who has never been quite comfortable with the other villagers or with the court, who doesn’t feel at home anywhere, and who has not much left to lose. (This justifies a wider variety of endings for him, of course — he’s more likely to be willing to run away from his home kingdom if he hasn’t got a wife and kids he’s leaving behind, say.)
5 thoughts on “LICK TREE. PURCHASE ANTLERS.”
Seriously, I’m sure the Infocom guys would have paid half a million dollars for such a tool.
TIE FACE TO SKY has a certain haunting quality to it.
I wound up making a modified spin-off of this extension to test NPC responses to instructions as well (as in, SNOW WHITE, TIE FACE TO SKY). It’s a little less generalizable than the original extension, but it’s catching even more infelicities (since covering a range of NPC commands is not always one’s first implementation priority).
Now that’s a tool!
Another idea — crowdsource it, the way some companies like Mahalo are trying to create better search engines results using human-power: