Re: CSS aims

On Thu, Jan 23, 2014 at 3:31 PM, Simon St.Laurent <simonstl@simonstl.com>wrote:

> On 1/23/14 2:18 PM, Brian Kardell wrote:
>
>> One of us is clearly talking past the other.  You keep attributing
>> something to me which I feel is the opposite of what I've said.
>> Please describe to me what you think is the decades of experience and
>> what that would mean to our way forward and how I am incorrectly
>> labeling that purist?
>>
>
> The decades of experience that you dismiss as "purist" is the:
>
>
>  "purist notion that there is a very important difference and that
>> presentation should be the realm of CSS."
>>
>
> It's not just an abstract principle - it's something that's done well for
> a long time.
>
>
This is actually exactly the opposite of what I perceive to have been said,
so let's please try to correct the misunderstanding somehow....

This is a long-standing (at least since 1996) view held by some that *all
aspects of presentation* should be expressed in CSS.  This is not my
opinion, it is what I call a purist opinion.  They disagree with your
notion that the principle is done well - I know this because they are
trying to remedy that and explain why/how it isn't done well.  More on that
below...





> Then you ask for:
>
>
>  when I need boxes and relationships that aren't in the DOM
>>
>
>
Here too - I think you are misattributing or misinterpreting something.
 I'm not inventing that - there are tons of proposals or even some things
already in CSS that do this.... Let's leave our own judgement of that on
the doorstep for a minute.

If you read the piece about Regions being considered harmful (and I expect
you did) the #1 issue Hakon raises is not that Regions can express
relationships with boxes not in the DOM - it's that doesn't *require* that
you create those boxes in CSS - Regions *lets* you use the DOM.  You can
verify that I am not misreading this intent because it is asked point-blank
right there on www-style[1].  Purists would (pronounced "are currently and
have been") attempt to block anything for layout that isn't *defined* in
CSS because that violates the core principle, or prime directive, or
whatever you want to call it.  It's not that they think Regions is too
radical, it's that they think it's not radical enough.

In any case - this thread isn't really about regions specifically it is
about whether this infinitely pure separation that is desired by some is
actually plausible or fitting with all of the other efforts we have in the
space (Web Components, <template>, etc)... I think not.  That's my
judgement.

I **absolutely** believe we can make progress further than where we are -
some middle ground where it is still simple, majority DOM but really really
handy ----  ::before and ::after generate presentational boxes from CSS
that are not in the DOM in the normal way - I find them to be wonderfully
useful additions.  I've very frequently wanted to select a fragment and
::wrap it or something, but couldn't so I have to go add a <div
class="whyamihere"> for no real reason.

If these "things" exist in any form (some already do) they require
explanation both in terms of how they work and the relationship to the
other things in the system (their role).

[1] - http://lists.w3.org/Archives/Public/www-style/2014Jan/0304.html





>  I suggest that you're asking for the wrong thing.  Boxes - certainly as
> expressed in the regions spec - may seem like what you need.  However, the
> shortcuts that spec takes in its rush to get there will cut you badly in
> the long run.  Relationships that aren't in the DOM are more complicated,
> but standardizing those seems like the wrong path.
>
>


> At least that's what my experience makes painfully clear, and I have too
> many scars from past cuts to demonstrate it.
>
> I hope that's clear enough.
>
> Thanks,
>
> --
> Simon St.Laurent
> http://simonstl.com/
>
>


-- 
Brian Kardell :: @briankardell :: hitchjs.com