Re: [whatwg] Default (informal) Style Sheet

At 5:38 PM +0200 UTC, on 4/7/07, Alfonso Martínez de Lizarrondo wrote:

> 2007/4/7, Sander Tekelenburg <<mailto:st@isoc.nl>st@isoc.nl>:
>
>> It seems extremely unlikely to me that every UA will in fact implement the
>> exact same default Style Sheet, even it it were mandated by the spec.
>> Therefore, in reality authors will not be able to rely on this and thus
>> speccing it, giving authors the impression they can rely on it, is likely to
>> be result in more problems for users, not less.
>
> So it comes down to this:
> You don't trust that the UA will implement the default Stylesheet, and so
>you say that every UA can do what they want because you propose that every
>web author include that CSS Zapper.

That's not at all what I said. I listed 4 specific objections initially. So
far people are responding to details only.

All I said in what you're responding to is that [1] it seems unlikely to me
that UAs will actually all use the same default Style Sheet and [2] users
will still have User Style Sheets, and that therefore the argument of
"ensuring Web pages will look the same across UAs" doesn't fly.

That doesn't mean I don't still question whether it is *desirable* at all to
try to make Web poages look the same across UAs. It doesn't mean I don't
still question that such a goal would even be achievable, because the same
presentation will not work well in different browsing environments (indenting
lists 40px is a waste of screen real estate on mobiles). And it doesn't mean
that I don't still think that speccing a default presentation will make it
harder for UAs to improve their default Style Sheet.

I *do* recognise the problem that currently most authors do not consider the
fact that UAs have default Style Sheets. So they end up serving CSS that
relies on that built-in Style Sheet, resulting in problems when that built-in
Style Sheet is different from what the author anticipated.

Authors can handle this reality in two ways
[1] when you want padding:0, explicitly state so instead of relying on the
fact that some UA already uses that as the default
[2] have your CSS start out with a 'CSS zapper'.

That second approach is easier for most authors, and is implementable for
authoring tools, which is why it seems the best approach to advocate.

> Some web authors will implement a CSS zapper to achieve the goal of having
>the same rendering in all the browsers

That would be a misguided argument. The goal of using a 'CSS zapper' is to
avoid relying on UAs default Style Sheets. I'm not claiming it would ensure
sites look the same across browsers. I in fact oppose that goal.

>, but too many web authors will focus
>just on the browsers that have the most usage

Maybe. I think advocating CSS authors include a 'CSS zapper' in their CSS is
a simple enough 'trick' that there is a reasonable chance authors will pick
up on it.

But assuming authors won't, I still don't see how the spec aiming to "ensure
Web pages will look the same in every browsing situation" could possibly have
better results, while I *do* see serious potential downsides to that approach.

> and new browsers will have to
>reverse engineer and copy those de facto standards that aren't documented
>anywhere or too many websites will look wrong.

Yes, that's the one argument that seems to make sense to me thus far. But it
should still be weighed against the objections.


-- 
Sander Tekelenburg
The Web Repair Initiative: <http://webrepair.org/>

Received on Saturday, 7 April 2007 18:10:57 UTC