<p>Probably a noobie question. What are the differences between an interface and a struct with function pointers? They look the same to me.</p>
<hr/>**评论：**<br/><br/>arrans: <pre><p>TL;DR the struct approach would result in spaghetti code, be more error prone, and require implementing logic that is essentially already handled by the compiler.</p>
<p>Although you could probably achieve almost the same outcome with both, the struct approach will end up being more cumbersome. In addition to having to piece together your function-pointer-holding structs rather than simply passing a value, it won't play nicely with all of the existing Go infrastructure (other packages, standard library...).</p>
<p>There will also be a decoupling of values from methods so you'd have to figure out, for example, some sort of closure approach whereby each function contains a pointer to the value with which it's associated. These are things that are handled by the language rather than the engineer so don't reinvent the wheel.</p>
<p>You also won't be able to use type switches with the struct approach.</p>
<p>As for what's happening under the hood, this isn't something I'm familiar with but would guess that it would be much less efficient even if negligibly so for most intents and purposes.</p></pre>chewxy: <pre><p>An interface is a struct with 2 fields. The first field is typically implemented as a pointer to the table of function pointers. The second field typically points to the data. </p>
<p>In a way, you are correct, that structurally they are similar. But semantically, interfaces are checked at compile time. This gives the compiler the ability to tell you that something is not right (for example, your types that do not fulfil an interface)</p></pre>
