Scoring on all exported symbols is overly punishing

[panva]

Consider the following module

export interface FooOptions {
  /**
   * Foo ...
   */
  foo: string
  /**
   * Bar ...
   */
  bar: string
}

/**
 * Foo Class Docs ...
 */
export class Foo {
  constructor(options: FooOptions) {
   // ...
  }

  // ...
}

Despite the class itself and every option being well documented and intellisense kicking in for new Foo({ ... }) this only yields 50% of documented exported symbols.

I would like to suggest that interfaces and type aliases be not considered in the scoring when it comes to exported symbols, or at the very least, have a much lower weight.

The alternative is that authors need to add rather pointless documentation such as

/**
 * Options for the {@link Foo} constructor
 */
export interface FooOptions {
  // ...
}

[crowlKats]

I would like to suggest that interfaces and type aliases be not considered in the scoring when it comes to exported symbols, or at the very least, have a much lower weight.

I dont agree with this; a lot of APIs relying mainly on having options available via an interface, or functions returning complex objects.

I think having to specify a comment in this scenario would be correct. It makes navigation much easier especially with a comment like you pointed out, since it would let you easily jump to the symbol that the options interface is for.

[panva]

I'm not in full disagreement with you, but, do you think the example above should be at 50% over the lack of a useless comment, which frankly, most of them will be because interfaces like these will always have a self-descriptive name?