W3C home > Mailing lists > Public > www-talk@w3.org > September to October 2000

RE: Connection: close (was: Re: Msg header)

From: Nic Ferrier <nferrier@tapsellferrier.co.uk>
Date: Sat, 02 Sep 2000 23:32:12 +0100
Message-Id: <s9b18e2c.037@tapsellferrier.co.uk>
To: www-talk@w3.org
>>> "Fish" <fish@infidels.org> 02-Sep-00 10:01:49 PM >>>
>> Another reason why the UA should pay more attention to the
"close"
>> value of the Connection header than the "keep-alive" value is that
if
>> the client keeps the persitent connection open there will be a
delay
>> whilst the server waits for timeout before bringing the
connection
>> down.
>So?

So that's not very user friendly - generally produces a delay in the
UA whilst it waits for the next response. This was Jims problem - he
was actually complaining about the performance of the system he was
using and narrowed it down to the connection header.


>> That's not very user friendly...
>Please explain how a user is *negatively* impacted by a client 
>keeping a persistent connection open to a server.

By the client running slow because it it dedicating time to the
handling of the connection.


>> One more reason for closing the connection in Jims 
>> case is that "Connection: keep-alive" is not a defined 
>> part of the HTTP/1.1 spec.
>So?

Errr... so the only acceptable understanding of the semantics of the
request is to close the connection (14.10)


>> The token keep-alive has no meaning. See section 
>> 14.10 of RFC2616.

>Wrong. See section 19.7.1 of RFC 2068.

Or section 19.6.2 of RFC2616. Yes - I understand about the HTTP/1.0
backwards compatability but the server is an HTTP/1.0 server. So
RFC2068:19.7 and RFC2616:19.6.2 don't apply.


>> It could be that what is being talked to by Jim is an 
>> HTTP/1.0 webserver connected to an HTTP/1.1 
>> servlet engine (or possibly vice  versa).
>Huh? You're making this up as you go, aren't you?

Actually this is exactly the problem. I have dealt with this problem
with Jim privately and also with the authors of the servlet engine in
question. It has come up more than once on Servlet-Interest.


>Now go lie down and sleep it off so when you wake up,
> these RFC thingies won't seem so "tricky" to you.

Curiously enough Jim's problems were caused by a proxy acting as part
of a test system.

There is a point to be made here - but you're not doing it. Obviously
you feel that you are far superiror to everyone else who corresponds
on this list and that your reading of the spec is correct. You also
seem to feel that abuse and rude words are a good way to make your
point.

When you have a point to make and when you can argue your point
sensibly please feel free to email me again. Untill then feel free to
send me lots of posts about your superirority and the alarming size of
your penis. 

Also feel free to explain to me how your brother is bigger than my
brother and how much ice cream you once ate in an afternoon.


Nic
Received on Saturday, 2 September 2000 18:26:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:24 GMT