Re: HTTP should be able to transfer part of a document

• From: Larry Masinter <masinter@parc.xerox.com>
• Date: Sat, 11 Mar 1995 16:15:43 PST
• To: DINGLE@ksvi.mff.cuni.cz
• Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
• Message-Id: <95Mar11.161545pst.2763@golden.parc.xerox.com>

Partial document retrieval based on the original URL is not consistent
with the semantics of HTTP URLs, where the data returned is not at all
required to be idempotent with the original data.

> 1) If a long transfer is interrupted, it's possible to resume the transfer
> without beginning all over again.  (This was a major motivation for the FSP
> protocol, which allows partial transfers; FTP does not.)

FTP has the same flaw: the file being FTPed could have changed without
your knowing it.

I suggest if you want to do this that (a) HTTP GET should return
(optionally, of course) a unique ID for the actual content. (MD5
signature, or at least name & date). (b) any attempt to GET_PARTIAL
take the ID rather than the rest of the URL.

 

Received on Saturday, 11 March 1995 16:20:23 UTC