Re: APP Meta-discussion

[Kyle Marvin <kmarvin@xxxxxxxxxx>]

On 7/20/05, Tim Bray <Tim.Bray@xxxxxxx> wrote:
> 
> <co-chair-mode>
> 

> 3. I think PaceCategorize is dead.  Lots of -1's, nobody was willing
> to stand up and say "yes, this is good".

Yes, as one of the original proponents I agree that it should be
withdrawn from consideration.  It is too narrow to include in the core
protocol and didn't necessarily handle all scenarios well enough even
if it was not.   I'd still like to see something here but agree that
this PACE is DOA, and will take the initiative to rethink it.

> 4. The ball on PaceBatch is still in play.  Unlike PaceCategorize,
> there are people who like the idea.  However, there are others who
> say "premature optimization".  Kyle says he'll do some numbers to
> help us ascertain what, in the real world, PaceBatch might buy us.

I'm able to put some focused time into PaceBatch testing this week,
and hope to have something to restart the conversation by the end of
this week or early next week.

> Although the Google people never addressed this thread directly, they
> made it clear they'd like:
> 1. Category tagging of entries in place without replacing the whole
> entry
> 2. Boxcar updates, a bunch of CRUD operations batched up in one message.

Thanks for summarizing, Tim.  This is accurate.

-- Kyle