- From: Ned Freed <NED@innosoft.com>
- Date: Fri, 09 Feb 1996 00:28:05 -0800 (PST)
- To: "Roy T. Fielding" <fielding@avron.ICS.UCI.EDU>
- Cc: Larry Masinter <masinter@parc.xerox.com>, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
> Putting a note in the spec about interoperability problems for URLs > greater than 255 characters long is reasonable, but the protocol itself > must not have any such limitations. There are plenty of ways to approach this that do not require putting a hard upper limit on URL size. For example, you could say that "all implementations are required to support URLs at least 255 characters long and implementations are encouraged to support URLs of as long a length as is feasible". Statements along these lines improve the situation to the extent that the specification will effectively deprecate implementations that don't support reasonably long URLs. Without such a statement there are no grounds for criticism. Any such "minimum maximum" requirement absolutely must be labelled for what it is, however There is an unfortunate tendency in the IETF for such bounds to end up being seen as simply "maximums", and I agree with Roy that this is undesirable. Ned
Received on Friday, 9 February 1996 00:41:47 UTC