Probably a few scripts would need to change, as there may be some kernel differences as MI indicated. And even though the classes are totally separate from the kernel, the interpreter does make some assumptions about what methods and properties are on each class. But I suspect they're fairly minor, given that I was able to "port" a trimmed down SQ5 to the LSL6 interpreter (but more differences may be exposed if you're trying to get the whole game working on LSL6 terp).
You'll see in the thread I linked that vocab.994 needs to tightly tied to the class hierarchy, as this provides a map of which properties are where in various core classes. That is, it contains information like "x is the 10th property on the Actor class". If this doesn't match exactly, nothing will really work. I forget what changed I made. I *think* I just grabbed the vocab.994 from LSL6 and used that, and then made sure that core classes (Feature, Prop, Sound, Ego, Room, Region and every other class that inherits from them etc...) has property lists that matched the LSL6 classes. Maybe I didn't even need to make any changes, I forget. At any rate, it should be easy to compare the SCI 1.1 template game core classes to those you get from decompiling SQ5, and see if there are any differences in the property lists (I added a "case" property to Feature and its subclasses, that's unrelated, you can leave that out). The SCI Companion documentation has diagrams that show the class hierarchy well.
I think there's also a version number in the resource.map file that would need to be manually "hex-edited" otherwise the interpreter will reject the game completely. It's near the beginning of the file, it should maybe be obvious by comparing the resource.map of SQ5, LSL6 and the SCI 1.1 template game. Oh, it actually looks like I hard-coded it to the LSL6 version number when writing out the SCI1.1 resource maps, so maybe there's nothing to do here except "rebuild resources", which is what will write it out. I hope I didn't break things for other games, yikes.
Finally, I think there were some global variables that needed to be correct, this thread might be useful:
http://sciprogramming.com/community/index.php?topic=1421.15The actual set of changes required is probably small. But when something isn't working, it can be very hard to find out why.