- From: Francesco Chemolli <kinkie@squid-cache.org>
- Date: Thu, 6 Nov 2014 19:03:15 +0100
- To: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Cc: Greg Wilkins <gregw@intalio.com>, HTTP Working Group <ietf-http-wg@w3.org>
>>My current guess is that the server is going to implement it's own
>>heuristic based priority mechanism that will use very little input from
>>client supplied priorities, specially not dynamically provided ones.
>>Hopefully I'm wrong and the current tree/weights will be useful to the
>>server side, but only experimentation will tell.
>
> +1
>
> I'm still of the opinion that no significant loss of performance would
> happen if the client just sent a single priority bit meaning:
>
> '0'
> The fetch is happening in a background context ie:
> * outside of scroll-regions viewport
> * in covered tab
> * in hidden window
> * screen-saver is running
> * from a batch-job
>
> '1'
> The fetch is happening in a foreground context ie:
> * in view
> * user is tapping his fingers
> * from a real-time transaction
>
> And I think the entire priority verbiage should be replaced by
> something like that and any more advanced priority communications/state
> relegated to extensions.
It's a bit extreme, but it's the kind of simplicity that pays the
biggest returns on implementation complexity.
I like this; can we discuss the idea a bit, and maybe straw-hat it?
--
Francesco Chemolli
Received on Thursday, 6 November 2014 18:04:05 UTC