- From: Russell O'Connor <roconnor@math.berkeley.edu>
- Date: Sat, 3 Feb 2001 15:33:38 -0800 (PST)
- To: W3C HTML <www-html@w3.org>
On Sat, 3 Feb 2001, Murray Macdonald wrote: > Russell, > > Thanks for a great substantiated reply. I guess that this would be > true if the included content does not change or have any conditional > requirements. How would an HTML client-side syntax allow for a single > included item not to be cached? I guess the included file would > require a pragma no cache tag. The problem of stale cached is not unique to this situation. Entities retirieved over HTTP come with an expires header that gives a suggested amount of time keep it in a cache. ... Anyway I'm far from an expert in this area. > Compatibility with existing UAs is still the best reason for SSIs. This is true. Adding anything new to the internet that isn't backwards compatable requires users installing new UAs. This holds for Flash, VRML, greek characters in HTML, etc., and will likely hold for client-side includes. People who make entites that use client-side includes would have to be aware of this. -- Russell O'Connor <http://www.math.berkeley.edu/~roconnor/> ``Paradoxically, a refusal to `put a monetary value on life' means that life is often undervalued.'' -- Artificial Intelligence: A Modern Approach
Received on Saturday, 3 February 2001 18:33:40 UTC