Re: [imap5] Designing a new replacement protocol for IMAP

[Adrien de Croy <[e-mail]>]

thanks for the pointer on BOSH.

<OT>

Some of the claims they make are patently false however.  Especially in 
http://xmpp.org/extensions/xep-0124.html Sec 4.

"Therefore, if over time the traffic to and from the client is balanced, 
bandwidth consumption will be about the same as if a standard TCP 
connection were being used"

this completely ignores the overhead of the HTTP request and HTTP 
response, which can be very large.  A protocol layered over BOSH which 
makes lots of small transmissions will have its bandwidth consumption 
greatly increased.  IME most transactions have an http overhead > 1KB.  
There are minimum requirements for headers which must be included in 
order to be valid HTTP.  So telnet over BOSH would increase bandwidth 
consumption by 500 times.

It's quite amazing this statement remains in that document.

</OT>


On 16/02/2012 11:53 a.m., Dave Cridland wrote:
> On Wed Feb 15 22:51:37 2012, Adrien de Croy wrote:
>> long polling is a hideous hack.  Proxies hate it.
>>
>> It's basically designing a system to provide a TCP over multiple HTTP 
>> over TCP connections.  Bloat to the extreme.
>>
>> I understand the reasons why it exist, due to the model of HTTP, but 
>> building more things on top of it heading in the wrong direction IMO.
>
> There are two options if you want to live in the web world - BOSH or 
> WebSockets.
>
> WebSocket support is *far* from universal, and BOSH works - and works 
> very well with proxies, despite whatever personal feelings they may have.
>
> Dave.

-- 
Adrien de Croy - WinGate Proxy Server - http://www.wingate.com
WinGate 7 is released! - http://www.wingate.com/getlatest/