W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1996

Re: About that Host: header....

From: Koen Holtman <koen@win.tue.nl>
Date: Fri, 22 Mar 1996 15:23:38 +0100 (MET)
Message-Id: <199603221423.OAA02021@wsooti11.win.tue.nl>
To: Harald.T.Alvestrand@uninett.no
Cc: jg@w3.org, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Harald.T.Alvestrand@uninett.no:
>
>JG's recent message <mid:9603160004.aa17589@paris.ics.uci.edu>
>seems to be uncontroversial with respect to what the host: header should
>look like if there is one, so I'll assume that's settled.

Just as a data point: I too think JG's recent message is
uncontroversial.

>The issues list has the text:
>
> Full URL must be accepted by server, may be sent by client.
>
>Since JG seems to think that uncontroversial, I'll assume that's settled too.

Me too.

>The questions before the WG seems to be:
>
>- Should full URLs be mandatory to send for all HTTP/1.1 clients when
>  they don't know that they are talking to a HTTP/1.0 server?

Making this mandatory is unacceptable for me.

>- Should the host: header be deleted from the spec?

No.

>- Are there less extreme positions that make sense?

Yes.  Make the Host header required if a 1.1 request line without a
host name is used.  This does not mean that we have to keep Host in
1.2.  As Jim said:

|We may be able to remove this lint in HTTP 1.2, if 1.1 really becomes
|ubiquitous.  But then again, we may not, if 1.0 servers don't die a
|timely death.


The arguments in favor of making the host name required in request
lines sent by 1.1 clients have a fatal flaw.  They start out with
noting (in Jim's words):

|The Host decision, if made where the current rough consensus I feel is,
|will accellarate the day when we must deploy a replacement protocol.

Then, they go on to argue that a switch to an incompatible replacement
protocol is very painful, so accelerating the day when the switch must
be made is bad.

However, the proposed alternative, making the host name required in
request lines sent by 1.1 clients, would mean making a painful switch
*right now*: all 1.0 servers have to be upgraded as soon as the first
1.1 client is released.

Surely, if accelerating a switch is bad, requiring a switch right now
is _extremely_ bad.  Not that I think that a switch to host names in
the request line will actually happen if we require it in 1.1:
browser authors will correctly predict that their customer support
departments would get swamped with bug reports (your new beta does not
work with site X!) if they would implement this part of 1.1.

Host is the lesser of two evils.

>                            Harald A

Koen.
Received on Friday, 22 March 1996 06:29:23 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:49 EDT