RE: Styling HTML placeholder attribute


[Tab Atkins Jr.:]
> 
> No, and Sylvain's similar assertion is similarly wrong.
> 
> Arguments from theoretical purity are interesting and all, but that's the
> lowest level in the hierarchy of constituents.  
> The most important thing is to figure out what kind of styling would 
> produce the desired effect for placeholder text, 

Our design and arguments are informed by a year of continuous user feedback, 
which indicated there were many desired effects. Stating we should not choose 
a common design on the basis of one *single* specific styling of a *single* 
part of a *single* input control state is not 'theoretical purity'. Nor is 
saying we shouldn't base our decision solely, or even mainly, on the need to 
work around specific limitations of this same specific styling. 

Suggesting we should instead consider fixing the opacity limitation that otherwise
makes a pseudo-element more convenient for grayed-out default styling is, in fact, 
entirely practical, as confirmed by Brad among others. 

> and then we reason 
> backwards from there to figure out what kind of properties and/or 
> selectors we need to achieve that effect in
> the best manner.  Trying to do it in reverse is architecture astronautics,

'Theoretical purity', 'architecture astronautics'...is there any reason to 
believe the pejorative dismissal of others' opinion will help us collaborate
on a solution? Or maybe that's not the goal?
 

> and often has bad results, as it is explicitly trying to reach a decision
> without reasoning about author/user/etc concerns.

We reached our decisions after reasoning about user concerns - some of which
I shared with you, most of it here -  and based on their requests e.g. 
using this control state to also provide visual progress feedback while they
fill in form (using border width and/or shadows, among other things).

> 
> Going from the proper way to deciding these things, it's reasonable to
> assert that partially-transparent text is a pretty good way to achieve the
> effect we want (text that always looks like a "washed-out" version of
> normal text).  We can then reason backwards from there and figure out what
> solution is best, based on how well it implement that styling and what
> additional kinds of problems the solution can incidentally solve.

No, thank you. I have no interest in 'reasoning' backwards from a one single 
use-case when user feedback suggests several others, no matter how 'proper' 
you will conveniently assert it to be. But since you will stubbornly persist 
in pretending your pet use-case to be the only one that matters and dismiss 
all others as a 'theoretical' waste of time I conclude this discussion to 
be a dead-end.
 
As for the color alpha discussion, I suggest a new thread.

> 
> ~TJ

Received on Thursday, 24 January 2013 20:22:42 UTC