Re: Next Web Focus

On Jan 25, 2014 7:29 AM, "Simon St.Laurent" <simonstl@simonstl.com> wrote:
>
> On 1/24/14 10:15 PM, Brian Kardell wrote:
>>
>>  > I thought this group was focused on polyfills and extending the
>>  > browser using the tools available within it.
>>
>> The aims of this group are, yes focused on advocating and discussing
>> polyfills and prollyfills - and advocacy that enables this, for
>> significant reasons and in ways described on extensiblewebmanifesto.org
>> <http://extensiblewebmanifesto.org>.  Very key among these is that it
>>
>> provides an evolutionary model forward rather than allowing a browser to
>> stymie efforts or prompting big bets on radical change which, if they
>> fail (and many do), stall forward momentum on the current platform.
>> Actually, i would say the majority of work has been applying this
>> advocacy in existing WGs to prioritize efforts in this direction as a
>> guiding philosophy.
>
>
> Nothing in the mission statement or the manifesto requires the kind of
derision for separation of concerns that has been a regular feature of this
group in the last few weeks.
>

I was going to let this sit, but actually, i want a record in the archive
for posterity lest someone find or link to this leaf.

There is not derision for SOC. There are repeated statements that we like
SOC, we want SOC.   There is also the illustration that it is not fully
separate now and the statement that there are many proposals attempting to
fix in giant leaps this by introducing significantly new abstractions and
doubt expressed by some that this is a positive step.  We will likely get
it wrong in committee - i am dubious that it is possible without stepwise
progress.  It's fine to disagree with that - in fact - I'd love to find a
proposal I could get behind without concern.. I can't.  Can you show me
one?

>
>> There is a significant amount of new magic being discussed in the CSS WG
>> aimed at advancing the goals I described in the post I put to the group.
>>
>> There are also  members of the WG attempting to explain the existing
>> magic and create a more extensible system where this sort of
>> experimentation can take place outside and follow a more evolutionary
>> path without a big bet.  Since there are wholly lacking primitives or
>> explanations necessary to provide the level of separation that they
>> seek, this is substantial.  One of these (Regions) is attempting to at
>> least explain the newly proposed magic in many of them. The majority of
>> the regions draft attempts to lay out the fundamental primitives that
>> can be used to explain the rest.  Debating the appropriate primitives is
>> fair game, but there has been little of that actually.  I'd love to see
>> more.  In any case, the one major issue that a lot of people seem to
>> have is that it recognizes something about the fact that the DOM is,
>> indeed, involved in rendering today - and that as such, in order to do
>> something practical today the only box that is pragmatically useful
>> until those other ones come about is the one that is there now.  This is
>> fine for many, but met with reproach by some who prefer something purely
>> more in line with their aims, which, as I say introduce some pretty
>> significantly new things at a high level.
>
>
> There's magic and then there's magic.  The Web already holds a lot of
magic.

It requires explanation in many cases that is still lacking in order to
allow realistic evolution of the platform.  Literally all that is requested
is to prioritize that work and introduce minimal new magic, and when you
do, explain it.  I've mentioned regions several times - clearly you think
it is abysmal.  That's fine too - but in order to understand where we go
from here (in terms of advocacy) it helps if we can explain why and agree
to something that doesn't have those problems.  "Not that" is less helpful
without answer to "ok, what then?".

Is it simply the fact that in addition to the CSS created regions you -can-
target an element?  I've explained why I view that as actually an advantage
in evolutionary terms, What if that were removed?  This is where I am stuck
hopelessly right now.  If you respond to nothing else, an answer to this
would be helpful.  You can see that Regions links to numerous proposals -
do you dislike them as well?  All of them are attempting to increase SOC,
the question is how and will it work.

Intelligent people can respectfully disagree on the above question - or - I
am hopeful that there is a good intersection of agreement worth pursuing.

"Pragmatically useful" in this conversation seems to be the polite way to
say "I have this tactical need and don't have much concern for its impact
on the larger strategy."

I've asked, fairly directly, if someone can point me to a proposal that
doesn't suffer from this perceived problem which we could focus on instead
and explained that I've not seen one and am dubious of its current
existence, and why I think that is so.  Again, I'm quite happy to be proven
wrong and very willing to redirect efforts - in fact - while I can't speak
for him, I can tell you that the author of Regions has shown similar
willingness for many months now.

>
>> It seemed significant to me and that we ought discuss in terms i thought
>> relevant.  That seems to have gotten a little off in the weeds.
>
>
> From my perspective, this thread went off in the weeds from the outset.
 I watched for a long time before deciding that it was time to point out
how strange it was.
>
I'm quite happy to not merely participate in an echo chamber.  Please
continue to contribute concerns.  If at all possible, examples, very
specific things you disagree with and rationale, or pointers to something
you think is better for us to focus on is helpful.

>
>>  > From my perspective, the very separation of concerns this derides as
>> a dead horse is what makes polyfills and browser extensions possible
>> without infinite tangles.
>>
>>  > Based on this message, though, it seems like the "next web"
>>  > perspective is something much more severe, leaning toward throwing
>>  > over the things that have worked in the past in favor of rapid
>>  > application development.
>>
>> I am beginning to worry that we have some significant communication
>> barrier going on here.
>
>
> No.  We have a values barrier here.
>
It's possible, but i think that the gap, if there is one isn't the chasm
you perceive it to be.  I think that at least is exacerbated by a
communication barrier.

>
>> My comments and aims, and it reads to me like
>> those of others as well are all around minimizing the severity of change
>
>
> Conversation about minimizing change doesn't help when the actual
layer-blurring proposal reaches directly to the heart of how we build web
applications and sites.  "It's just one tiny little change", but there's a
much larger impact on the overall ecosystem.
>
> You asked earlier for examples in code, but the problem, like most
architectural problems, isn't actually in code.  The problem is in the
integration and social layers above it.
>
>
>> to the standards in platform at once and create an environment in which
>> we can prove out and improve before we invest years or even decades into
>> significantly new very high level abstractions in standards.  Where new
>> magic is necessary, explain it such that continued work and improvement
>> can harness significant mindshare and be proven out.  I really don't see
>> how it could be taken otherwise.
>
>
> Rushing for the new magic while polyfills still need a lot of love and
attention seems like a major mistake to me.
>
> I get the logic of ever more ever faster to get mindshare and shiny and
racing to meet the expectations of developers who just want to build apps
and happen to be stuck using web tools. I'd prefer, though, that extensions
of the Web recognize the foundations that made them possible.
>
> The work described in the Extensible Web Manifesto has only begun.  The
virtuous cycle of API development through polyfills still needs a serious
jumpstart.  I agree that it's the way forward, and would much rather see
focus on building a polyfill ecosystem than pushes for changes to the basic
structures of HTML/CSS/JS/DOM interaction.
>
> Browsers certainly are exposing more functionality through APIs, but are
regions really a "low-level capability"?  It seems unlikely to me.
>
Excellent: why? More specifically, the topic of this thread was intended to
be about the many other proposals it attempts to provide explanation for.
Are these problematic as well?  Each of them is attempting to introduce
increased SOC with new magic.  I feel, and I think our group should feel,
that this should require explanation.  Do you?  Is there something which
does seem like a better primitive in there?

>>  > Is that really the purpose of this group?
>>  >
>> I hope i have clarified my original intent in posting this to the
>> group.  If not, I'm happy to have a more rapid exchange directly if you
>> feel that would be  more helpful and we can always loop it back in if we
>> come to any form of resolution.
>
>
> For now I plan to return to observer status, and will work outside of
this group to promote what I thought it was trying to do.  I've been doing
that anyway, for a very long time.
>
> (I'll still have an article for Wednesday, and will let you all know when
it goes up.)
>
I really do look forward to reading it Simon and I'm glad to have you
promoting some stuff.  I really hope that you don't remain entirely silent,
more participation and discussion is better.  Again, I'm quite happy to
have this conversation privately, or publish a dialog.  Whatever makes you
comfortable.

>
> Thanks,
> --
> Simon St.Laurent
> http://simonstl.com/
>