berichtnews: snipped-for-privacy@e10g2000vbe.googlegroups.com...
New Yaaawk Ci-t-t-t-ty ?:-)
...Jim Thompson
berichtnews: snipped-for-privacy@e10g2000vbe.googlegroups.com...
New Yaaawk Ci-t-t-t-ty ?:-)
...Jim Thompson
berichtnews: snipped-for-privacy@e10g2000vbe.googlegroups= .com...
program
be
output
really
to
10FxxxRepeating myself, FTP is more than sufficient to the task. It recovers gracefully from line disconnects and user suspends. (picks up where it left off) =20 Bittorrent is actually a bit different approach, with the reliability as an add on to an inherently flaky (for large file transfer) http protocol.=20 .
There's no reason why HTTP should be any more flaky than FTP. Both handle resuming a download in exactly the same way, i.e. the client tells the server to skip the first N bytes of the file.
Nah, the 'burbs. Real NYC folks don't apologize for jumping in.
Cheers
Phil Hobbs
handle
No. They don't. Do go read up on the protocols. Http is more like udp. Ftp is a multifile wrapper around tcp. Big differences. .
I don't need to read up on them, I'm already familiar with both protocols.
HTTP uses TCP. If you want to claim that it's "more like" UDP, you'll need to elaborate.
No idea what you mean by "multifile".
The biggest difference between the two is that FTP uses separate control and data connections. The control connection persists so long as you are logged in, and is used to initiate transfers. A separate data connection is established for each transfer. To resume a download, the client sends a REST command whose argument is the length of the partial file, causing the subsequent RETR command to start transferring data from the specified offset.
HTTP is stateless; there is no persistent connection. The client opens a connection, and sends a request, and the server returns the data on the same connection. The request may include a Range: header to indicate that only a portion of the data should be returned. To resume a download, the client specifies a range whose starting position is the length of the partial file.
handle
protocols.
need
Plenty of implementations with user interfaces allowing/tracking many files at once, from many directories, as the user requests. Was around long before the web.
athe
Gotta love sockets.
It is the statelessness that makes it inappropriate. Besides it is not clear that ranges are required. 14.35.1 begins saying that it is not required of all servers or all clients, but more complete reading does not support the assertion.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required