Re: HTTP Caching Model?

At 12:32 PM 12/13/94, Daniel W. Connolly wrote:
>I'm not sure what you're suggesting here.

What I'm suggesting is that competent production people will work with
whatever the real world users have their.  Incompetent production people
will assume a perfect world where everyone has infinite bandwidth, and
bug-free clients.

What I'm suggesting is that the statement

>>>| User-Agent shouldn't affect the retuned data. (The fact that it
>>>| does is a wart that we'll have to deal with somehow.)

is unreasonable given the lack of a perfect world.

Now - if we had URCs and version control .... then clients could return
explicitly different URLs for different browsers, and have the caches do
the right thing, but I'm probably dreaming.

>I can see that real world nasty problems require real world nasty
>solutions.
>But as far as a spec, shouldn't we use the categorical imperative?
>i.e. what if everybody did that?

If everyone did that, then more documents would display reasonably on more
browsers, but caching would be much harder.  I'll gamble on upgrading poor
caches more than I'll gamble on upgrading poor clients, because there are a
couple of orders of magnitude fewer.

>If everybody customized their documents on a per-user-agent basis, and
>caching proxies don't take the User-Agent: header into account in
>their cache keys, then things will be broken.
>The question is: where do we assign the fault?

Assigning the fault doesnt help - once the bad clients are out there, it
does a production group no good at all if they assume they dont exist.
>
>If we say that the proxy was broken for not using User-Agent as a
>cache key, then we're saying that cached objects can never be shared
>accross clients with different user agents. I don't think we want
>that.

Just an idea ..... Allow for a User-Agent field in the mime header of the
returned document. If a server wants to recognise the User-Agent field,
then it should return a User-Agent field to say this document is only valid
for that User-Agent.

>So I'm suggesting that serving up different documents for different
>User-Agents should be a protocol violation.
>>Of course .... if none of the browsers had bugs, and all did a good job of
>>presentation, then designers wouldnt need to server up multiple versions of
>>files.
>A spec tells you what happens when everybody plays by the rules. When
>there are bugs, all bets are off, and you do what you must.

And I'm saying that when the clients dont play by the rules, then the
servers cant. So this particular rule should be modified.

- Mitra

=======================================================================
Mitra                                                    mitra@path.net
Internet Consulting                                       (415)488-0944
<http://www.path.net/mitra>                           fax (415)488-0988

Received on Tuesday, 13 December 1994 16:03:56 UTC