NodeList

I tried using this for a unit-test, and ran into some issues with e.g. `childNodes`, which in regular DOM is a [NodeList](https://developer.mozilla.org/en-US/docs/Web/API/NodeList) and not just an `Array`, which is missing, for example, the `keys()`, `values()` and `entries()` iterator methods.

I can foresee issues as well with e.g. `children` being a plain `Array` created by filtering `childNodes` - whereas, in regular DOM, mutating the `childNodes` of the parent that generated the `children` collection, would cause the iterator to reflect those changes.

Maybe this is "by design", favoring simplicity over correctness?

If so, that isn't clear from the "Project Goals" stated in the README right now - it doesn't seem to imply that this deviates from DOM in terms of more than features.

Lacking features is a matter of scope - but incompatibilities probably ought to be designated in the documentation?

README の Project Goals から始め、childNodes と children が現在どのように生成されているかを調査します。NodeList の keys()、values()、entries() メソッドおよび報告されている mutation/liveness の期待値と動作を比較します。完了の条件は、互換性の範囲について明確な判断を下し、文書化された相違点を示すか、範囲を限定した実装変更とテストを行うことです。

索引モデルが issue の本文から書いたものです。