W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2015

Re: Request header in response

From: Florian Best <http@florianbest.de>
Date: Mon, 16 Nov 2015 18:50:03 +0100
To: ietf-http-wg@w3.org
Cc: Dave Wain <dave.wain@ntlworld.com>
Message-ID: <564A174B.1020805@florianbest.de>
If you don't do it with an ID you cannot differentiate between two
unsafe requests which have the same headers.
And of course there is a lot more traffic if you would send all the
header back.
Well, I think you could implement this by yourself if you need it.
Implementing it in HTTP seems not necessary and not backwards compatible.
And you should have a look at HTTP/2.

It would be nice If you would answer to the mailinglist instead of me

Am 16.11.2015 um 17:59 schrieb Dave Wain:
> Florian,
> Good idea, you could use a request-id. Reminds me of when I
> invented web sessions back in 95/96.
> In this case however it would just be easier to include the actual
> request because the id would not be guaranteed to be unique
> and caching would be difficult.
> Good discussion though ;)
> DW
> -----Original Message-----
> Sent: 15 November 2015 20:00
> To: Dave Wain <dave.wain@ntlworld.com>
> Cc: ietf-http-wg@w3.org
> Subject: Re: Request header in response
> Okay, for this you just need a unique identifier for the message. e.g.
> one header "Message-Id: 12345"  in the request and response.
> Am 15.11.2015 um 20:50 schrieb Dave Wain:
>> But to match the response to the request requires (some of) the 
>> request header to be part of the response header.
>> For example, if requests go via a satellite phone but the response 
>> comes via a satellite broadcast channel, there would be no way to 
>> de-multiplex the responses.
>> Regards
>> Dave
>> -----Original Message-----
>> Sent: 15 November 2015 16:18
>> To: dave.wain@ntlworld.com
>> Cc: ietf-http-wg@w3.org
>> Subject: Re: Request header in response
>> Hello Dave,
>> I would like to understand in which way you think that HTTP depends on 
>> TCP/IP and how it wouldn't if the response contains the request headers?
>> IMHO the underlying communication protocol has nothing to do with it.
>> As long as you can parse the message syntax and make the connection 
>> management you are independent of the underlying transport layer:
>> You could even transfer HTTP messages via UDP, UNIX sockets (as I 
>> sometimes do), or whatever you want.
>> Am 15.11.2015 um 06:05 schrieb Dave Wain:
>>> Dear All,
>>> If the request header (or most of it) is included in the response 
>>> header, HTTP would become its own transport layer, independent of 
>>> TCP/IP.
>>> This would be especially useful for remote locations.
>>> DW
>>> Note: This message is private and confidential and hence must be 
>>> received without interception or distortion by the intended 
>>> recipients only. Permission to use the information explicitly must 
>>> come from the sender (and recipients).
>>> My personal web site and alternate contact details are at:
>>> http://homepage.ntlworld.com/dave.wain/.
Received on Monday, 16 November 2015 17:50:32 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 11:11:40 UTC