Reading their list of the advantages of modular animation, I feel like that system is more suited to platform and rpg games, where there are a lot of characters engaged in a lot of action.
In adventure games where only one or two characters tend to walk anywhere, and most of the animations are about expressing attitudes or emotions, I don't think it would offer as much.
Ah, but think of the opportunities for making more complex animation from smaller sprites! Giving someone a different stance (held frame) while their head still talks, talking while walking (Bud Tucker style - he even chewed gum while walking, something that not many adventure game heroes are able to do!), or putting different facial expressions on the characters themselves while not bloating out your game with thousands of redundant pixels.
Close up shots could still benefit from this limited animation, and be done ten times more elegantly than the current method.
You could change the character's clothes independantly of their mood, they could dance and still be able to talk.
The advantages of this system isn't just for Rayman-style animation, but a whole spectrum of animation-saving processes. Who needs to add in seperate "blink" frames and mess around with custom talk animations and unwieldy views for Sierra portrait talking, when you can just composite all you need in Spriter, and do, say
vGraham.Mood (MOOD_ANGRY);
cGraham.Say ("I am very angry!");
And you add new #defines and moods directly on the artist's side, and not on the programmer's side where they have to use up new Views to change something as insignificant as an eyebrow. Heck, if the lip sync tech in AGS was improved, you could leave it to the lip sync dude to note which lines are which mood, and change it appropriately (coincidentally, a lip sync editor in AGS would be badass as hell). Done totally procedurally, no changes or custom Say command with complex animation code required.
I'm just blue-sky thinking at the moment, but the possibilities are endless! Most of these could probably be done in a few hundred lines of code and several sprite slots, but Spriter may prove to be much more elegant in practice. The prospect of easy-to-do reusable limited animation intrigues me!