Is it a bad idea to couple action and animation?

Question

I'm making an ARPG (like Legend of Zelda but with more action), and this question raised during the development:

Should action and animation be coupled? Or should they be independent?

With this I mean, the process of executing an action (eg. Attacking) should be:

1. Input
2. Start Animation
3. Frames 1-3 are transitioning, do nothing
4. Frames 4-7 are attack moves, do the attack action during these (searching enemies in front of the character, pick the closest and deal the damage)
5. Frames 7-10 are transitioning, do nothing again
6. Animation finished, release input again.

This way, it's like the action is triggered by the frames, as events on the animation itself.

If it helps, I'm making the game on an Entity/Component System.

Should it be done this way? Why yes/no? What are the advantages/disadvantages?

Answer 1

You really don't have much choice. Ignoring some (IMO) terrible other options, you can either:

a) couple your behavior to specific animations - so artists don't make one attack animation, they make several that your behavior plays at the appropriate times

or

b) have animations emit events and export state variables that game logic is driven from - requires lots of effort put into tools to make the workflow nice, but lets gameplay be driven by art and design far more easily which is usually a very good thing

If you have the time and expertise to develop the tools, I'd go with the second option every time, but small indie/hobby/old-school games do tend to lean more towards the former.