Re: Procedural (non-technical) point about freezing the cat and hat combinators before they've even been defined (was Re: Shadow DOM: Hat and Cat -- if that's your real name.)

On 2/5/14 2:11 PM, Tab Atkins Jr. wrote:
> I don't know how much you've followed Blink's development in the past
> year

Honestly, not that much.  I've been busy with other things.

> nowadays, Blink practice is to just drop a UseCounter on it, wait a
> few weeks for data to come in, and then make a relatively informed
> decision.

Sure.  This part I'm aware of and it's much appreciated.  This approach 
works great for "is this API called?" or "is this API called with 
certain arguments" or "is this syntactic construct used?".  It's less 
great for dependencies that can't be easily determined in code like "is 
this web page depending on this event firing on this particular node?".

> Plus, we've been explicitly positioning things (through Polymer) in a
> way that should hopefully let us be much more aggressive about pushing
> changes.  Polymer (or the other major component-wrapping libraries)
> can adjust to a lot of changes by just tweaking things internally, in
> a way that doesn't break authors.

I don't know enough about Polymer to say much about how this affects the 
risk of compat issues with the problems I listed, unfortunately...

Does polymer abstract away the styling of the shadowroot and CSS 
inheritance, for example?  I would think not, since polyfilling CSS 
inheritance is kinda hard...

-Boris

Received on Wednesday, 5 February 2014 19:17:36 UTC