Hello,
The "hard coded" nature of the species, superclass, name, and -info-
selectors kind of bothered me until I realized that if one went only by the
script file format, there's nothing "hard coded" about them. I began
thinking about this again just now, and am wondering what sort of
flexibilty among selectors, their ID's, and their placements could be
expected from an interpreter.
I'm looking at a script from a 0.685 game, just in case any of this is
wrong for other versions. I don't think it is, though.
FreeSCI, and I assume the original Sierra interpreter as well, requires
that the first four selectors in an object or class be species, superclass,
name, and -info-, all in that order. There's nothing that I can see
philosophically wrong with having them listed some other way, however,
since class definitions include the property ID list. Going even further,
there's nothing philosophically wrong with giving these selectors
nonstandard ID's since VOCAB.997 can be used to lookup selectors with
special meanings to the interpreter.
I'm not saying that FreeSCI should allow this, but it does make me ask the
question of what other special selectors are there that don't get located
through VOCAB.997 and the property ID lists? AFAIK, those are the only four
in FreeSCI. Back in July, Lars mentioned that the Sierra interpreter has
hard coded selectors in event handling. Which are those, and what's
important about them: their property ID's, their offsets in the class
defition, or both? Is it fair to make any assumptions about them when
VOCAB.997 is missing?
Cheers,
Ravi.
