Re: Multipart Support Proposal

> Do we really want to overload XMLHttpRequest with "comet-style" 
> functionality? XMLHttpRequest is already quite complex.
A good point. Like I said before, multipart support is complex, and the 
value is not super high. I would like to see it, but I am not going to put 
forth a lot of effort to defend it. However, I believe most comet developers 
will continue to lean on XHR even if HTML5 server-sent events become more 
pervasive, and the extra connection and pipeling control are extremely 
valuable tools for that with minimal API addition. Pipelining control in 
particular has significant value even beyond "comet-style" long-lived 
connections, and as I responded Maciej, could help us through the current 
pipelining impasse.
> I'd personally  (and I think the company I represent would agree) much 
> rather go with the  server-sent events proposal in HTML 5.
Obviously, your company was the first (and only one so far?) to implement it 
;). No offense intended, it just that for a number of reasons, comet 
developers have not been universally eager to use utilize this approach, 
many preferring more XHR centric approaches where there is more direct 
control of the API (no need to go through DOM elements) and the 
communication format.
The idea of this proposal is standardize (with appropriate fixes) Mozilla's 
existing multipart support, which could be a lot easier for comet developers 
to utilize.
Thanks,
Kris

Received on Tuesday, 19 February 2008 16:25:47 UTC