Re: Advocacy / prioritization

I'd add "first load bootstrapping" for ServiceWorker (née:
NavigationController), which would enable polyfills to truly control
resource loading in the page, starting from the page's first load (unlike
the way it is defined today).

As an example, ServiceWorker/NavigationController was mentioned in the past
as a way to polyfill CSP. Without it handling the first load, that is not
something you can rely on (since first page load users will remain
vulnerable).

Ragarding tap.js, assuming the concept takes off, combining ServiceWorker
scripts in should be something we think about as well. (for URL rewriting
polyfills that are used across sites)



On Thu, Sep 19, 2013 at 8:35 PM, Brian Kardell <bkardell@gmail.com> wrote:

>
> On Sep 19, 2013 2:07 PM, "Clint Hill" <clint.hill@gmail.com> wrote:
> >
> > It's hard to really articulate 1 thing, as much as I could mention a few
> things that are all really related within a major problem I face day to
> day. The issue is within the development of "widgets" or "components" or
> whatever you call a re-usable feature. The low-level infrastructure for
> these things is missing in a lot of ways, however the crux of the problem
> is in the transport and delivery of code/assets/resources for these things
> and it's not being addressed (at least not that I'm aware). I believe the
> Web Components Intro (http://www.w3.org/TR/components-intro/) captures
> the low-level infrastructure the best. I would say from that the list that
> is important to me is:
> >
> > HTML Import
> > Custom Elements
> > Templates
> >
> > And these are all moving along as specs. However the core issue not
> addressed by these is in the delivery of these things with respect to
> performance and efficacy. I believe the POC tap.js is part of the answer (
> http://bkardell.github.io/tap/) and has the right central idea to
> resolving performance and efficiency.
> >
> > If somehow all these things could be called 1 thing then I would say
> that would be a priority to me. These topics would get me a lot in my day
> to day work.
> >
> >
> > On Thu, Sep 19, 2013 at 5:24 AM, Brian Kardell <bkardell@gmail.com>
> wrote:
> >>
> >> One of the stated charter items for us, and key in the extensible web
> manifesto is advocating prioritization of necessary low level features with
> standards bodies and browsers.  Might be interesting if we could discuss
> which things we think are worth advocating and make a concerted effort on a
> few of them.
> >>
> >> Anyone care to propose things that are currently pretty lacking at a
> low level and problematic?
> >
> >
> Tap is heavily under rework, i will share as i get further along next week.
> I think the stack for it uncovers (in current unpublished form) some
> pieces i would like to see soon: basic promises actually native (nice work
> on that today), messagechannel ports - mozilla is the last major hurdle for
> basic support there but they are working on it.  setImmediate.  Async
> promise based apis and further interop on a few key things: idb,
> localstorage, xhr.  Prollyfilly or polyfilling all of these currently takes
> a fair bit of code is doable at all.
>
> But...i am asking if there are others - maybe even fundamentally lacking
> ones which we can assemble into some kinda list.
>