>Could we discuss the benefits/drawbacks of this algorithm? Oh, Brian, there you go again actually creating solutions instead of arguing about philosophy and who understands the problem better. :-) I'm no proxy maven, but this solution appears workable for both servers and end clients. Given that a client that wants to get more of an aborted download can just as easily create a "Request-Range: bytes=" as it can a URL as proposed, I believe that the optional HTTP header from the client is the preferable solution than the overly-extended URL. Basically, users mess up URLs all the time, but I feel safe in letting client software handle headers. If both schemes give the same result, headers will be more reliable and less prone to user error. --Paul Hoffman --Proper PublishingReceived on Monday, 13 November 1995 09:50:30 UTC
This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:42:56 UTC