Re: Server-side roles in the HTTP

> In that case, I think you want the "Accept-Encoding" header, not TE. (Make
sure
> you're looking at a fresh copy of the spec.)  The issue is of
content-encoding
> vs. transfer-encoding.

No, I was thinking of transfer-encoding, which can be used for
compression-type services.

Content-Encoding would be interesting, but potentially it really isn't as
user-transparent as most of the other facilities I'm talking about. Will
think about it...