rish3 - 11:26 pm on May 5, 2013 (gmt 0)
Anyhow, from my perspective, both Google, as well as many others on this forum are confusing "real proxies", like squid, with these "web based proxies".
All of the RFC's pertaining to proxies are talking about real proxies. Real proxies don't take HTTP requests like a normal web server (with the exception of "transparent mode"). They use ICP, HTCP, etc. The one exception is "transparent mode", where they DO NOT DO CRAP LIKE REWRITING URLS. THE PROXIED PAGES ARE 100% VERBATIM.
Somewhere, someone at google's DMCA team, and apparently here as well, "web based proxies" somehow got equated with "real proxies".
There are no RFC's for a web-based proxy that rewrites the content. However, a sensible web-based proxy would take measures to ensure that it doesn't create duplicate content...like a <meta ROBOTS> tag for example.