Re: [Tools-discuss] Asynchronous submission

[Jennifer Richards <[e-mail]>]

On 2023-05-30 11:48 a.m., Carsten Bormann wrote:
> If the concern is that the server might be loaded by the refreshes, the refresh wait should really be extended based on the server queue size, not on the status of the draft’s processing.

We don't currently have queue details available where they'd be needed 
to make this happen. I am also not sure that there's much of a win vs a 
simple backoff. After the first 10-15 seconds of waiting, it's already a 
"long" wait and the perceived impact of an extra 15 seconds (or 
whatever) after the draft was actually processed is likely to be minimal.

The motivation for using status API is to do better than the current 
method of reloading the entire page before there's an update. The reload 
causes visual glitches on at least some browsers that'd be nice to 
avoid. (A rewrite of the page to be properly dynamic is planned, right 
now just trying to make things usable enough until we get to that.)

Thanks,

jennifer