Theory/opinion question: What are good RESTful archive URLs?

Suppose I have a weblog application with RESTful resources accessible
at the following URLs:

GET /posts
GET /posts/1
GET /categories/
GET /categories/1
GET /categories/1/posts

As implemented the semantics of /posts and /categories/1/posts are
such that they show “current” posts, with current defined to be a
rolling window of a certain duration.

My question is this: what do you think is the best way to represent
the archives of older posts as RESTful resources?

We could do this:

GET /posts/archives
GET /categories/1/posts/archives

or this:

GET /archives/posts
GET /archives/categories/1/posts

or this:

GET /archived_posts/
GET /categories/1/archived_posts

or, undoubtedly, others. What do you think is the most defensible
approach?

Note: if you don’t like the idea of having GET /posts return only
“current” posts then you can perfectly well invert my question. Let
GET /posts retrieve the full set of all posts known to the system.
What then is a good RESTful representation that limits results to
“current” posts?

Thanks in advance for your thoughts,

Sven

Quoting S. [email protected]:

As implemented the semantics of /posts and /categories/1/posts are
such that they show “current” posts, with current defined to be a
rolling window of a certain duration.

GET /posts/1.html # “current” posts
GET /posts/1.xml # all posts

The former is presumably a browser with a human in front of it. The
latter is
another app that can define “current” for itself it it wants. Same
applies to
other non-human readable formats, e.g., json.

Just my $0.02USD,
Jeffrey