9 character word recognition?

In the Inform 7 manual (3.1, section on Descriptions) that “a small limitation here is that probably only the first 9 letters of each word are read from the player’s command.”

Why is it probably? What would make it read more than 9? At first, I thought “Wow, this is odd” and the example given (superhero, superheroine) seems like a great example of why you wouldn’t want that limitation in place. But then as I was playing around, I realized I never actually ran into an issue with anything I was doing.

Why 9 in particular? Since the manual says this will probably happen, what are the cases where it won’t?

I think you can change this number when compiling to Glulx with the line:

Use DICT_WORD_SIZE of .

This doesn’t work with Z-Code though!

Ah ha. That got me the phrasing I need. Thanks. I searched on that and found reference to it in Adaptive Prose in the Recipe Book. I’m learning a lot, but the Inform manual sure does make you ping-pong back and forth!

I think the manual should remove the “probably”. It sounds like it’s a definite unless you add this very specific phrasing.

The Z-machine uses a variable-width encoding; punctuation takes more bits. So Z5/8 can recognize “aldebaran” but not “ping-pong”.

In Glulx it’s all fixed-width.

Interesting context! Thanks. That actually should probably go in the manual in some way shape or form. Although I guess so far I haven’t even had to worry about “Z5/8” or “Glulx.” Someone also mentioned “Z-Code” and I started searching for those phrases and mostly found stuff that I wasn’t sure how it mattered at all.

Wow, I had no idea you could adjust this. Thanks for the tip!

(I did in fact run into this problem when I wanted to distinguish between “think about Luxembourg” and “think about Luxembourgish” in Ollie Ollie Oxen Free. I wrote around it instead.)

This tends to be a bigger problem when writing German stories, since we do have more (and sometimes longer) compound words than in English…

It has always been possible to distinguish between longer words by referring back to the input buffer. There’s an exercise in the I6 DM4 somewhere.