Tue, 30 May 1995 00:48:24 +0500

I have a question for the various browser authors out there.

I'm currently implementing some of my web replication research, but
I'm not sure how browsers currently handle redirections. If a browser
requests a URL, and my server returns a redirection to another
machine, do most browsers cache this information? If the links on that
new page are relative do most browsers append the new, redirected
machine to form a URL? Or do most browsers continue to append the old
machine to the local file to form the URL? Are these redirections
commonly cached across sessions (only applicable for browsers that
cache information across sessions, clearly).

Let me present a scenario to make my questions clear.

User A connects to Server X to request the following URL -

File 1 has been replicated on Server Y, and Server Y is closer to User
A. Therefore server X returns a redirect message - http://Y/file1

User A now retrieves that file from Server Y. In that document was a
reference to a local file, of the form <a href="file2">link</a> When
the user tries to retrieve this file, what will the browser do?
Presumably it will try to retrieve http://Y/file2 right? what if only
file1 has been replicated on server Y? I assume this will fail, since
I presume the browser will not attempt to retrieve
http://X/file2. Does this imply that I should stay away from local
references? That when server X replicates the file onto server Y that
it should replace all local references with the appropriate absolute

James Gwertzman.