W3C home > Mailing lists > Public > www-style@w3.org > August 2010

Re: [css21][css3][svg] SVG and unit-less length values

From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Date: Sat, 14 Aug 2010 13:53:31 +0200
To: public-fx@w3.org, www-style@w3.org
Message-Id: <201008141353.31814.Dr.O.Hoffmann@gmx.de>
Sylvain Galineau:

>
> Breaking pages that use such stylesheets may be acceptable to a user like
> yourself who knows how to override font sizes using advanced browser
> settings or a user stylesheet and knows what CSS is and how to turn it off.
>
> None of which is true for 99% of web users (at least).
>
(The reason for maybe 30% of them could be, that they still use MSIE -
even I did not find out up to now how to switch off stupid
stylesheets with this browser and how to make pages accessible -
sure, users of such a browser are delivered to the incompentence
of authors of such stylesheets without any help from the browser -
fortunately there are alternatives ;o)

As already mentioned, it does not break any page or even styling -
it does just what the author noted, that has to be done - it is the
decision of the author. If a simple program like a browser tries to be
more clever than the author or a recommendation tries to predict
and avoid all stupidities an author can provide, this will result in
a Cascading Stupidity Spiral and not in Cascading Style Sheets.
The result of such an approach can be already seen in the HTML5
draft - because browsers did something wrong, authors do something
wrong and finally the author of the draft recommends to implement 
something even more broken and unintuitive, to cover all these
previous nonesense. Indeed it is hard to believe, that many authors
will ever be able to follow this path of bugs, mistakes, misunderstandings,
errors and pessimisation to understand, what they have to do to get
a predictable result.
Related to CSS units, if you specify that if an authors notes '1cm' he
will get in the average say '0.9cm', but sometimes '1.2cm' and 
occasionally '0.5cm' this is a clear indication for any reader
of the recommendation of a stupid format with an urgent requirement
for improvement or replacement.


> > It causes only problems for CSS for (X)HTML style sheets, if they are
> > not worth to be displayed anyway.
>
> And who decides whether they are 'worth' to be displayed ? 

The author - there is always the choice to provide something that is
not broken. I think, from all my style-sheet there appeared only a
few problem at all due to surprising changes from CSS2.0 to CSS2.1
or due to common bugs in major browsers, however, once detected
it was always simple to fix the problem.

> Is it reasonable 
> for a user's banking web site to be unusable on his new computer with a big
> fancy new display because a working group he never heard of decided that
> the site's stylesheet was no longer worth displaying using current display
> technology ? What's more important: the stylesheet's political correctness
> or the user ?

Such a bank should have enough money to provide an accessible
web-site, managed by persons who know what they do. 
If they not even manage this, this is a clear indication, that they will
not be able to manage other jobs either ;o)
Therefore this will never appear for say much more than one or
too days for a bank trustworthy enough  to administrate your money.

It is a help for the user - it is a clear indication for the question, whether 
such a bank really cares about its clients or only about the
shareholder value ;o)

This has nothing to do with 'political correctness' of the format or
the working group. It is a question of integrity.
If a technical recommendation contains obviously stupid nonsense,
it is not usable for serious application. This is an major problem for
all W3C formats and for internet content at all - to get away from the
impression, that this only amateurish fun stuff compared to something,
that can be really used for serious applications.
And looking on the quality of browsers and web sites, we are currently
far away from a situation, that a W3C format can be really used
straight forward to provide some primary content, not only additional
fun content apart from serious business.

>
> (Mental image of message box saying 'Sorry. This stylesheet was not
> worthy')
>
> >The font-size is never under control
> > of the author due to minimal font-sizes. If the style sheets depend on
> > wrong assumptions or on a wrong/awkward choice of units, nothing
> > can save such a style sheet - it was already broken, when the author
> > created it.
>
> It may be broken in the solution space that you prefer. But if that
> solution space assumes it is OK to break millions of web pages for
> hundreds of millions of users as punishment for the past transgressions
> of a few web designers, 

This has nothing to do with punishment, it means only to take seriously,
what authors do, if they want to have '1cm' they should get '1cm' and
nothing else. This finally avoids the impression, that they are cheated
always by faulty browsers, which force them to write '1.2cm' if they need
'1cm'. And such impression results finally in the conclusion, that 
implementors or recommendation editors are either completely
incompetent to solve even such simple tasks or sadistic to force
authors to do stupid things to get a meaningful result.

Authors will always do nonsense, but if the behaviour of implementations
it at least straight forward and predictable, they have a pretty good chance
to fix the bugs pretty soon. Obviously if they have to work through a 
cascade of historical bugs this will finally frustrate them to fix bugs at
all and to learn how to do it right, because such a history is irrelevant
for the current problem to provide some content in a simple and
straight forward format.


> it will lose out to those solutions that 
> produce no such outcome. And the browser that does choose to punish users
> for all the 'unworthy' but otherwise valid CSS that worked fine for years
> may well lose out to those browsers that elect to preserve the layout
> integrity of existing content as pixel density increases. A process which,
> of course, implies a non-trivial amount of cross-browser interop pain.
>

If this cascade of bugs gets too complex, frustration of authors will
increase even more as well as the desire for a new format without
this problem - or simply to use inaccessible raster graphics to avoid
all this historical nonsense.
Many people will still know the disaster with MSIE6, longing for years
and the enormous  efforts of many authors to work around all these
bugs just for this browser to provide something accessible even for
those people, who still used this outdated software.
I think, most of them are pretty tired of those stupid complication
and a desire for a straight forward format that really works.
This will be never the case if you already get 0.5cm to 2cm if you
only write 1cm ;o) This is a basic and simple task - and if it turns
out, that CSS does not even manage this, CSS finally failed and
is broken, what is even much more worse than fixing bugs of browsers
or in stylesheets - this happens for many years continuously with any
new browser version.

> This is not an optimal course of action, to say the least.

Well, is there an optimal course at all?
To have a historical cascade of bugs to follow is no real option for
new authors.
To fix stupid bugs in already existing pages may hurt some people
for a short time - but finally it will help them to solve their problems in
a predictable and straight forward way.
To continue a historical cascade of bugs will hurt all authors using
the format now and in the future - with no hope to avoid all this
nonsense some time at all.

>
> We can in fact 'save' such stylesheets. 

No you can't, this never worked and over all these years authors
had always the work to follow the cascade of stupid bugs instead
of fixing their own bugs.

> Most importantly, if we can 
> preserve users' experience of existing content as hardware improves then it
> is quite hard to justify solutions that do not, given the number of users
> and the amount of content involved.
>

Well my basic experience is, that CSS in MSIE at the first place and
in other browsers in several details as well causes a lot of problems
and work, because it practically never happens, what is expected.
And if the behaviour is more or less ok for the audience, there is typically
an advanced author behind it, knowing a lot of bugs and workarounds,
to get something predictable nevertheless. 
For years new authors had not just to learn the meaning of CSS
recommendations, they had to learn much more stuff and stupid
hacks to work around the cascade of historical bugs - or well they
had to say 'forget it, switch off CSS files for MSIE or use raster
graphics'.



> > It is the task of authors to choose units properly for CSS+(X)HTML,
> > because for this combination this is essential to get something
> > predictable with different font-sizes and viewport sizes.
>
> It always was their task. But that doesn't fix nor convert the content
> that is out there. 

Sure, no one prevents them replacing stupid stylesheet with
useful stylesheet. Many change their stylesheets anyway every
one or two years. They just have to do it 'right' the next time,
what can be pretty easy without having a cascade of bugs
on one's back.

> Browser vendors are not in the business of punishing 
> millions of users to educate thousands of designers on the mistakes of
> their past work.

As already mentioned, it is not a question of punishing or guilt.
Nobody is perfect and there will be always some bugs around.
Fixing them instead of insisting, that they have some use or
application, is the only hope for authors and implementors to
get rid of a cascading history of bugs.
It is a question of taking authors seriously.
Self-determination and confidence instead of  paternalism -
not really new ideas at least since  the 'Age of Enlightenment'.
I cannot see a need for an 'Age of Darkening' currently ;o)

Once the format is corrupted with too much stupidity, there is no
hope for authors anymore to get it right - at least with this
format, it is burned. And than there is indeed someone guilty -
the editors, that recommended such a cascade of bugs as
a recommendation.

>
> Not the ones I know, at any rate.

Olaf
Received on Saturday, 14 August 2010 11:54:06 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:30 GMT