Re: [CSS21] body { padding: 8px; } HTML40 sample stylesheet

On Sat, 26 Jul 2003, Tom Broxton wrote:

> Watching users who are not so familiar with the web, struggling  to
> understand how to use it and even what hyperlinks are, has made me
> believe quite strongly that some consistency in how web pages work is
> important to users.

Fair point - link appearance (and functionality, of course) can be
regarded as being part of user interface, so to say. But existing sample
style sheets don't help here. They set links underlined, which is
_debatable_ (and Opera takes another root), and the very common _author_
opinion on that (after all, suppressing link underlining is probably the
most common first use of CSS) really reflects _user_ opinions to some
extent. I think authors should not prevent links from being underlined
but neither do I think that a CSS recommendation should say that browser
vendors are _encouraged_ to make links underline by default. Regarding
colors, we have the fantasy settings in the CSS 1 sample sheet, nothing in
the CSS 2 sample sheet. And nothing for :hover in either of them.
Regarding the common default link color schemes, they're really bad;
the only justification for not fixing them is that bad practice
(debatably) becomes good practice when it becomes very common.

A sample style sheet cannot help in such issues. If it recommends the
common practice, it effectively tells that the Web should forever be
coupled with some poor decisions made years ago and that any browser that
tries to offer something better by default should be "fixed".

> I think it is also helpful to authors because it means they are not
> having to reinvent the wheel every time they create a document.

Sorry, but it won't help here. The sample style sheets, to the extent that
they describe existing practices or say something that some browser
programmer might take seriously, are useful reading to learn what to
_watch out for_. For example, an author should know that clueless browsers
may display address elements in italics by default, and this is even
sort-of recommended by the W3C - thus, some evasive action might be
needed. But authors should surely not _rely on_ the sample sheets as being
used in browsers. It's sufficient to say that _no_ browser actually
applies the the principle that the sample style sheet be used as browser
default style sheet.

> > Surely. But that relates to the general principles, not specific
> > styling,
> > and it should be said in HTML specifications.
>
> And surely CSS is the best language that we have for describing what
> those principals should be?

No, English is the only feasible language for that. The _principles_ of
rendering should be described in prose, irrespectively of any style
sheets. The CSS "language" could then be used to express specific
implementations of the principles - but it would be really foolish to make
any particular style sheet an _example_ that implementors are encouraged
to use as such, or even as a starting point.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Saturday, 26 July 2003 07:05:37 UTC