Granted about the relative RESTfulness. However, the default behavior of most web servers is to ignore the query string if the requested URL corresponds to a file (the cached resource) and not a program. This technique is useful if you want Nginx or Apache to deliver the content from disk and would prefer to avoid getting your hands dirty with URL rewrite rules.

Why is your first URI “less cacheable” than the second? They’re both just identifiers for resources after all. There’s nothing inherently “more RESTful” about the second than the first.

Jim