Archy is currently a single component in a larger system, but judging from announcements from the Raskin center, the name is intended to encompass the entire operating environment. Currently, it is something like a very idiomatic text editor extensible in python. Its aims are to produce a rich browsing and editing environment (including other media than text), within a single flexible and extensible environment. In this way it looks toward being a sort of GNU Emacs, but eschews the various modes that characterize it, emphasizing uniform behavior of keyboard commands regardless of the subject matter being viewed.
One interesting design feature of Archy is that it is orthogonally persistent: there is no "save" function. Everything you edit in Archy is persisted automatically, along with an unlimited "undo" buffer, so any change can be reverted. There is also no filesystem in Archy: documents of all types are created and joined by simple separator glyphs, editable like any other character. There is a simple "leap" interface to navigate documents, similar to incremental search, and more exotic mechanisms such as a ZoomingUserInterface (ZUI) are under development.
The release of Archy as of May 2005 is something of a proof of concept -- an "alpha" release that lives up to that designation. However, with full-time paid developers working from well-established (if not always accepted) principles, there should be a clear development roadmap. RCHI tends to play its cards rather close however (Eric Raymond would call them very much a "cathedral"), so the roadmap isn't quite as public as it should be.
Having tried Archy, I came away really disappointed in the feature set, and in several ways disgusted by its gratuitous break with well-ingrained standards such as key repeat. The cult of personality that Raskin still posthumously enjoys appears to allow for very little flexibility in interpreting, let alone questioning his vision. Criticism tends to be brushed away as mere misunderstanding of some grand vision (possibly the barrier to criticism is high, requiring a scholarly rebuttal instead of impressions and anecdote). This does not bode well for Archy, and sadly not for Raskin's overall goals. As a veteran Emacs user, I find quite a lot laudable about powerful keyboard control, and am more than happy to see a project that focuses on it. I just don't think Archy is the answer. -- ChuckAdams