I hate to tear open a wound, but it seems to me that two important browser
vendors have yet to buy into Shadow DOM.  It's currently listed by
Microsoft as "under consideration" but the sense I get is that the signal
isn't very positive right now.  Firefox is planning to move forward, Blink
has it unprefixed.

Things like document.register can be polyfilled fairly well and without too
much crazy.  If imports is controversial or we determine that we need more
experimentation to figure out "what's down there" in terms of other systems
like modules or fetch - we can do a lot of those experiments outside any
browser implementation too and use it to lead discussions.  I am all for
that, especially if we can lead the way in getting vendors to cooperate on
the polyfills and make some efforts to find future safe ways to do this.

But Shadow DOM - this is a different story.  It might not be a fundamental
primitive or DNA level thing, but it's well down there and actually
impossible to polyfill accurately and it is dark, dark magic requiring lots
of code to even fake it reasonably well.  There's a real risk there is that
the fidelity could actually cause problems when you jump to native too, I

There seems to be a pretty large split in sentiment on Shadow DOM, or
perceived sentiment from developers.  From my perspective, a whole lot of
people tell me that they find Shadow DOM one of the most compelling pieces
of custom elements and without it, they're holding off.  Another thing they
tell me that frustrates them is that this makes it hard to share custom
elements - should they assume a Shadow DOM or not.

With Mozilla's post the other day[1] this has opened up a whole lot of new
conversations on my part and the preeminent question seems to be whether
there will be a positive signal from Apple or Microsoft or whether we need
to consider that as good as vapor for now.  For a lot of orgs,
consideration of switching to custom element and their plan for the next
few years is probably affected, as well as the state of the landscape and
where we will be shaping it.

With this in mind, I'm asking if anyone is willing to tip their hand at all
- even to the effect that "if we get two interoperable, unprefixed
versions, we will follow"... Any information I think is helpful - and
asking the question at least might move the conversation forward again (I

1 - https://hacks.mozilla.org/2014/12/mozilla-and-web-components/

Brian Kardell :: @briankardell :: hitchjs.com

Reply via email to