RE: Re HTTP2: caching and copyright

Kevin Hoadley (
Mon, 11 Jan 1993 16:27:37 +0000 (GMT)

>>> Note that servers shouln't cache documents with restricted readership since
>>> each server don't know the restrictions to apply. This requires a further
>>> header to identify such documents as being unsuitable for general caching:
>>> Distribution: restricted | unrestricted

There are a number of reasons why it might not be sensible
to cache a document, eg in my local server I have a page
entitled "Current state of the network". This is a smart doc,
refreshed everytime it is accessed. Caching this is just plain
silly. Thus a general mechanism (separate from access restrictions)
is needed to prevent caching where it is unsuitable.
Given that a general mechanism is also needed for cache timeouts,
it seems sensilble to amalgamate the two, ie documents that should
not be cached simply have their cache timeouts (TTL's) set to zero.
Note that within the PROTOCOL, cache timing information should
not be embedded within documents, for the simple reason that in the
future such information will be needed for non-text bodyparts,
ie my network status page might become a graphic, but it still
should not be cached. Note that this does not stop any server
IMPLEMENTATION from reading caching info embedded into in HTML,
merely that the protocol (the bits on the wire) should be able to
seperate the two.

>It would be nice in some circumstances to define the readership groups for
>situations where a server could apply group membership information to
>restrict readership.
>This says that the document can be given to anyone in psl and anyone in the
>kbpd subgroup of incl. You can make these names correspond to your
>organisation. The maintenance of these readership groups is outside
>the scope of the HTTP2 protocol.
>Local servers shouldn't cache documents including this
>header unless they "understand" the specified readership groups
>and can apply the same membership tests.

How does a server know if it "understands" readership groups ?
How do I know if my Dave Raggett is the same as your Dave Raggett ?
This information cannot be passed between servers; we would need
a global naming scheme - X.500 distinguished names won't do
as X.500 is not commonly enough used, e-mail addresses probably
aren't good enough (at a quick count I can be reached through
about 20 addresses at 4 institutes in 2 countries, does this
mean that my membership or otherwise of a readership group
depends on where I am logged on ?). We would also need a global
authentication scheme ...
This comes back to the idea that the only documents that should
be cached are those whose read access is totally unlimited. Any
other form of read access (ie, billed access to private data,
logged access, restricted access) should require talking to
the original server, not to some cached copy.

Kevin Hoadley