[whatwg] [html5] Rendering of interactive content

2009/2/9 Smylers <Smylers at stripey.com>

> Giovanni Campagna writes:
>
> > So the whole rendering section is just for implementors and authors
> > should act if no default style sheet is present
>
> No; the section is also for authors, in that it advises them of how
> content is expected to be rendered in mainstream graphical browsers.
>
> > or worse, if it could be everything, like a inline-block <div> or blue
> > <table>,
>
> Indeed it _could_ be anything, but there's no reason for authors to be
> worried about that; if a (non-malicious) user-agent does something
> different from that expected by HTML 5 then it presumably has done so
> intentionally and it wouldn't necessarily benefit users for authors to
> attempt to combat it.

The fact is that could lead to mixtures (part of UA stylesheets and part of
author style sheets) caused by the cascade mechanism that cannot integrate
themselves.

> so that the author should set all supported properties to initial or

> the HTML5 "expected" value?

Not necessarily.

> That is: I, author, want consistent rendering on all plaforms and

> browser:

I think that's where you're going wrong.  Platforms themselves aren't

consistent -- in things like screen sizes, resolutions, distance between

screen and audience, whether they are interactive, what input devices

are available, pagination.


>From the same CSS, you cannot get different results in different browsers or
platforms (unless the browser are not complying with CSS). Content may be
scaled, wrapped, clipped, overflowed, but it will always be the same. A
mobile browser should be different than a resized Firefox window. (Actually,
IE in Windows Mobile / CE is like IE in Windows XP / Vista, just with a
smaller viewport. The same applies for Safari in the iPhone or in Mac OS).
In addition, you can have digital pens also in desktop PC, but you expect
that your browser renders in the same way.

>
>
> If an esoteric platform chooses to divert from the expected HTML 5
> rendering then it's likely because that platform knows more about what
> rendering is appropriate for that platform that you do.
>
I'm not sure the UA knows better than me how to render my web site, because
this would mean it knows how to render every website in the whole world,
better than its own author. It may know better than me how to render
specific parts of it (ie widgets), but not all of it.


> > I import the HTML5 style sheet inside author ones.
>
> That's a very parochial view.
>
> In mainstream graphical browsers such importing would be redundant,
> because they'll have the expected behaviour anyway.

Can I assume they'll have the expected behaviour? No I cannot, because HTML5
explictly says that UAs may render in every way they like.

>
> In esoteric user-agents you cannot know that the HTML 5 defaults provide
> a better user experience than those chosen by the user-agents'
> developers.
>
But I can know that it cannot provide a worse user experience, because, as I
said, given a CSS and a media type (screen - interactive - visual), the
rendering is always the same. Maybe scaled, wrapped, overflowed, but always
the same. (yellow is still yellow, overflow: scroll still produces scrolling
mechanisms, display:table renders a table, width:auto is like width:100% in
non-floated non-positioned block-level elements, etc.)


> > I, implementor, want to provide backward-compatible rendering for
> > those author that didn't follow rule 1), I import HTML5 style sheet
> > inside UA defaults.  In both case, a downloadable stylesheet would be
> > much appreciated.
>
> I think a downloadable style-sheet is inevitable!

Then let's wait for Ian to write one.

>
> Smylers
>
Giovanni
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.whatwg.org/pipermail/whatwg-whatwg.org/attachments/20090209/53898cd7/attachment.htm>

Received on Monday, 9 February 2009 07:51:22 UTC