Playing with CSS to create a "different browser"

[Jukka K. Korpela]

Some time ago in this group, someone suggested that we should develop a
"different" user style sheet to demonstrate what a user style sheet or a
browser style sheet _could_ do. I guess the idea was that the style would be
rather different from common browser defaults, yet sensible and potentially
useful.

Partly for that reason, partly to demonstrate the idea that web page
rendering _could_ be different from what people are used to, I wrote a
smallish style sheet
http://www.cs.tut.fi/~jkorpela/styles/cool.css
and a sample document for demonstrating it:
http://www.cs.tut.fi/~jkorpela/www/testel.html
(Using Firefox, just access the latter URL and select "A cool style sheet"
from the style sheet list in the View menu.)

The ideas include
a) not using underlining and colors for links but arrow prefixes, as in many
encyclopedias, but with link text highlighting on mouseover
b) using special background color rather than bolding for <strong>
c) indicating form structure better
d) using dashes as list item markers instead of bullets
e) "literary paragraphs".

The point is that everyone using a sufficiently advanced browser, like
Firefox 2, can effectively create her or his own browser as far as default
rendering is concerned. (Well, there's the technical limitation that style
sheet selection does not work across pages on Firefox, but surely someone
could develop a fix to that.)

Are such ideas worth developing? I think I must admit that I'm particularly
interested in opposing the HTML 5 ideas that aim at making current browser
behavior the standard, despite the fact that many of the typical
presentational features that we are now familiar with are really results of
poor and hasty design over a decade ago. To demonstrate why the default
rendering should not be frozen as a "standard", it's probably easier to
_show_ how things could be. (Of course, a new default rendering need not
please all; the very idea is allowing diversity.)

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

[Bergamot]

Jukka K. Korpela wrote:

>
I think I must admit that I'm particularly
interested in opposing the HTML 5 ideas that aim at making current browser
behavior the standard,

I have just started reading about this and have not yet formed an
opinion. I'd like to get more info before deciding.

despite the fact that many of the typical
presentational features that we are now familiar with are really results of
poor and hasty design over a decade ago.

With the exception of links, I find many of the browser default styles
follow conventions that have been used in word processors for decades.

Your test page:

http://www.cs.tut.fi/~jkorpela/www/testel.html
(Using Firefox, just access the latter URL and select "A cool style sheet"

doesn't mention anything specific that has a poor and hasty design. Can
you cite an example or 2?

BTW, the elements you gave alternate styles to (links, bullet markers,
form fields) didn't really strike me as improvements over current
browser defaults, though I realize you were just demonstrating that they
_could_ be different, not how they _should_ be different.

To demonstrate why the default
rendering should not be frozen as a "standard", it's probably easier to
_show_ how things could be.

As long as both authors and users are free to override browser defaults
with CSS (as your page clearly demonstrates), does it matter so much
what the specific browser defaults are?

I think that having agreement between browser makers is generally A Good
Thing. It certainly does make life easier for the rest of us. Whether
some of these things should be standards is questionable, but for now I
think I'll keep an open mind and continue reading about HTML5.

--
Berg

[Jukka K. Korpela]

Scripsit Bergamot:

With the exception of links, I find many of the browser default styles
follow conventions that have been used in word processors for decades.

For about two decades, and I'd call them defaults rather than conventions.
The defaults have been used so widely, because it is often difficult to
change them, or at least programs don't make it easy and obvious to change
them.

Compare this with the century-long conventions of printed texts. During the
last twenty years or so, there has been a clear divide between quality
publications and word processor class documents. The latter class includes a
large bulk of desktop publishing products and even many products from the
print industry.

Creators of graphic web browsers, starting from Mosaic, made them follow the
word processor paradigm without thinking much. At first, this was even worse
than in word processing, since there were no ways to change the
presentation, except by refusing to use logical markup. If you really didn't
like the idea of gross font size for headings and equal margins above and
below a heading, you more or less had to use <font ...instead of <h1>,
<h2>, etc. If you didn't like "engineering paragraphs" with empty lines
between them, you just had to use <brand not <p>.

CSS was supposed to change this, but authors already had bad habits.
Besides, the paradigm was already there. Authors were used to seeing word
processor "typography", so they imitated it or - more often - just let
browsers use it and concentrated on other matters.

Now it seems that such "conventions" might be regarded as established and
might be regarded as existing browser practice that shall be declared
Standard,

Your test page:

- -

doesn't mention anything specific that has a poor and hasty design.
Can you cite an example or 2?

It's more or less implicit in the style itself, but here's a brief summary
of common flaws in common defaults:
1) No restriction on line length (text column width).
2) "Engineering paragraphs" as opposite to "literary paragraphs".
3) Equal vertical spacing above and below headings.
4) High-level headings too large in font size, low-level headings
ridiculously small.
5) Serif font for headings and tables.
6) Monospace font for textarea.
7) No borders for tables.
8) Link presentation that is unsuitable (for largely opposite reasons) both
for inline links (in copy text) and for lists of links.
9) Too much indentation for lists.
10) Bullets as list markers.

I could add
11) The same style for print as for screen
but that's an issue I did not address much in my stylesheet, for simplicity.

As long as both authors and users are free to override browser
defaults with CSS (as your page clearly demonstrates), does it matter
so much what the specific browser defaults are?

Certainly, especially if the defaults are made a Standard. If something is
clearly unsuitable and inconsistent, why should it be made the default?
Browser defaults should constitute a consistent, useful style that is a
genuine alternative to using any author or user style sheet. For example,
when a user encounters a page that is thoroughly messed up stylistically
(foolishly small font size, grossly insufficient color contrasts, etc.),
then disabling author styles should give a reasonable rendering if the
markup is at least half-decent. And it often does, but the rendering need
not be simplistic along the Mosaic tradition.

I think that having agreement between browser makers is generally A
Good Thing. It certainly does make life easier for the rest of us.

Browser wars were not nice, but I'm not enthusiastic about an eventual
cartel either. After all, if we want browsers to be similar, we should ask
Microsoft and the Mozilla community agree on a turning Firefox into IE 8.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

[Bergamot]

Jukka K. Korpela wrote:

Scripsit Bergamot:

>With the exception of links, I find many of the browser default styles
follow conventions that have been used in word processors for decades.

Compare this with the century-long conventions of printed texts.

I didn't say word processors were necessarily correct or wise in their
defaults, just that the browser defaults seem to correspond in many
ways. You made it sound like browser makers pulled everything out of
thin air.

Now it seems that such "conventions" might be regarded as established and
might be regarded as existing browser practice that shall be declared
Standard,

here's a brief summary
of common flaws in common defaults:
8) Link presentation that is unsuitable (for largely opposite reasons) both
for inline links (in copy text) and for lists of links.

I'll agree that some of those defaults aren't great, but I have to
comment on this item in particular. In your test stylesheet you chose to
style inline links like those "in an encyclopedia" using an arrow
graphic instead of underlines and different coloring. I found this to be
most unsatisfactory on screen because links were no longer obvious.

Don't forget to consider the difference between print and web media, or
how people read on the web compared to a printed page. Your suggestion
may be fine for certain types of web documents or in a print stylesheet,
but I think not for the average web page.

9) Too much indentation for lists.
10) Bullets as list markers.

I personally don't care one way or another on how much indent or what
marker is used, though consistency would be nice. What I do care about
is marker positioning related to the text, especially the positioning of
graphic markers. This can vary greatly between browsers as well as at
different text sizes. I would be happy if there were some standards in
that area.

>As long as both authors and users are free to override browser
defaults with CSS (as your page clearly demonstrates), does it matter
so much what the specific browser defaults are?

disabling author styles should give a reasonable rendering if the
markup is at least half-decent. And it often does, but the rendering need
not be simplistic along the Mosaic tradition.

I disable author styles rather often, mostly for their foolish type
size, color choices or other things that make reading difficult. Maybe
I'm just accustomed to the simplified rendering, but I don't expect the
result to be print quality, I just want to read something quickly then
be on my way.

Maybe my expectations are low, but maybe yours are a bit high. All in
all, though, after considering everything you've said on the subject, I
think that making some of the existing common browser defaults
standards, like heading sizes and spacing, would be dumb. But other
things, like list marker positioning, are worth pursuing IMO.

>I think that having agreement between browser makers is generally A
Good Thing. It certainly does make life easier for the rest of us.

if we want browsers to be similar, we should ask
Microsoft and the Mozilla community agree on a turning Firefox into IE 8.

There is more to a browser than just how it renders a web page. Besides,
MS would never go for this idea. ;)

--
Berg