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

Re: [CSS21] 4.3.2 Lengths (where absolute size matters)

From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Date: Wed, 15 Dec 2010 17:47:57 +0100
To: www-style@w3.org
Message-Id: <201012151747.57871.Dr.O.Hoffmann@gmx.de>
Anton Prowse:
...
>> To obfuscate the meaning and to frustrate use cases for such
>> units is simply not explainable. I discussed it with other authors
>> and what I got are shaking heads about such an obfuscation.

>No doubt the WG would be interested to hear about the use cases.
...

Well, not sure, looking on previous discussions and the inconsistency
introduced in the current draft, whether they are interested in use
cases for absolute units or not. However, this is another issue than
the inconsistency, therefore it is better to discuss this independently.
Therefore I start a new comment.

There are many applications in education, science, techniques
and trade for absolute units.
They are often related to the desire to present an image of an
object in original size or to present a map or a technical drawing
with a known scaling factor. There are only a few meaningful
use cases for absolute units related to a font size however. For
most applications of font sizes CSS has much better approaches
than px or absolute units.

Presentation of images or drawings in original size helps the
user to compare unkown objects with objects of known size,
for example the monitor itself or the own hand etc.

In education often one cannot really expect that users,
for example young children or people with limited imagination
about scaling factors and rescaled objects, have the spontaneous
capability to have an intuitive feeling about the real size, even if the
author provides a scale within the image to indicate the
scaling factor one has to apply to get the true size of the
object from the image. This is especially confusing for many (most)
people, if the scaling factor is in the range of 0.5 to 2, because
this is close enough to the original size to be confused.
Unfortunately the error introduced by the reference pixel concept
is often in this range.
It is a common problem of our time, that especially children 
have much more contact with images and graphical representations
of objects than they have seen such objects in real life or have
the chance to put their hands on it. Therefore they need help
to orient themselves in real life with relyable information about
their real environment. Scaled images or those with unkown
scaling factors increase the phenomena of estrangement from
reality. Even adults with a lot of contact to the 'electronic twilight zone'
of our time can have a tendency of loosing contact to real life -
one indicaction for example could be, that they try to define a
centimeter as a multiple of pixels ;o)

The application in education and science is therefore to be able
to present images of objects in original size or with exactly
known scaling factors on monitors of PCs, notebooks, netbooks,
mobile phone displays.
Often such exponates will not be printed on a sheet of paper,
however this may currently already work better than the presentation
on a screen (according to a simple test, a browser like 
WebKit fails even for this application and is therefore not usable
for printing of absolute units as well).

For example a text like
'The animation shows an optical atom-atom collision.
The width of the image corresponds to 20 atomic units,
this means it is presented with a scaling factor of 
ten millions.' is much more informative, especially for
users not familiar with atomic units, than just 
'The animations shows an optical atom-atom collision.
The width of the image corresponds to 20 atomic units.'.
Therefore it is a big advantage for the reader, if the
author provides the image with a size in absolute units
and a defined scaling factor. 

The application for maps and technical drawings is obvious.
Technical drawings are often printed, but taking into account,
that monitors meanwhile can have a relative good qualtity and
that it is simpler and cheaper to present something in colors 
or interaction/animation with it, it is not necessarily the case, 
that all drawings are printed.
To get/check sizes with a measuring stick this requires a proper
presentation of absolute units as well.
Of course, with the common use of GPS and the upcoming
Galileo satellite navigation and the idea to add orientation
capabilities of devices to SVG it gets more and more interesting
to be able to provide maps with predictible scaling factors
and author and user defined magnifications.

Applications in techniques and trade are obvious as well.
Just this month I was looking for macro flash lights and I
would have been very interested in getting images or drawings
in true size of these objects from the companies, which
produce them - unfortunately it was not possible to get
this on my screen, therefore it was necessary to travel
to a shop (nasty in winter) to get the impression/information
from the real object. The problems are similar to those
described for education, but with the consequence, that
companies with the capability to provide representations
of products in true size have a chance to be less annoying
and more attractive for people. But to do due this, the presentation
on the screen has to be relyable, else people get the impression,
that those companies are cheating.

In trade the applications are obvious as well. 
In the past years obviously many people started to buy
things in online shops without having seen the objects,
they are buying, before they have them at home.
On the other hand, many people think, that size matters
for many products they want to buy and they want to
compare directly with other porducts or things, they already
have or are offered in other shops.
It would a big advantage to be able to present images 
and drawings of products in original size.
For example for (compact) digital cameras, mobile phones,
notebooks, netbooks, dildos, lingerie, projectors, books, 
indoor plants etc many people think, that absolute size
matters and a presentation of wrong size would be
interpreted as cheating, therefore such applications
will be only seen, if browsers manage to present absolute
sizes on screens of computers, mobile phone etc - and
this can only happen, if there are absolute sizes available
in common formats like SVG and CSS+(X)HTML.

Such screens typically have all a similar viewing distance,
therefore there is no problem for accessibility to have
absolute units in images available for such devices.
For pinters, absolute units are especially important
for technical drawings, maps and the indicated 
true size presentations for education and sciences.
I think, product information form companies can
be relevant as well. I think, information from online-shops
will be typically seen on the screen and not printed, therefore
for them it is less important to be printable.

All of these applications are not much related to a font size
or something like this, more to graphics and images.
However, even for font sizes there is a good application.
If a user has the option to set the (minimal) font size for
presentation in an (!important) user-stylesheet or in
preferences for screens and for printing, the same
size can be used for quite different devices, high resolution CRTs,
medium resolution TFTs for computers, notebooks and netbooks, 
smaller mobile phone displays and for printing. 
According to my experiments for all of them the viewing distance
is almost the same. Chosing the same absolute font size for all
of them saves a lot of time instead of finding out the resolution
and setting it individually for each device with pixel based units. 

Stylesheets for printing from the author are no useful application
for font sizes neither in absolute units nor in pixel based units.
However, as already mentioned, if the user decides about the
font size, absolute units are useful. And if there is a publisher
of a book or a teacher printing text for a seminar, those people
have to decide for the readers about the font size before printing.
Unfortunately there is no accessibility approach for the readers
in this case, therefore careful decisions  are required from
these publishers, there is no other chance to prevent them from
doing nonsense than education, no automatism. 


Obviously presentations with projectors are problematic, because
the viewing distance can be quite different. On the other hand,
such devices are typically used for talks and presentations and
the author prepares more image like slides especially for this
purpose instead of using the 'scrolling approach' from (X)HTML
and multi purpose documents, therefore SVG user coordinates 
and 100% width and height with a viewBox is much more 
useful for this application than absolute or pixel based units. 
Because there is a CSS draft for transformations as well, 
it could be a good idea to introduce such a concept of 
user-coordinate units and a viewBox for
CSS+(X)HTML as well for such applications. 
This helps much more than the obfuscation of absolute
units as in the current CSS2.1 draft.

Another problematic case are TVs, they can have a
similar size in pixels than computers screen, however the
viewing distance can be quite different. This case is not
covered by the current obfuscation of absolute
units as in the current CSS2.1 draft as well.
For this, the only chance I can see it to use device
dependent style sheets as well as for some other
more exotic devices, that are currently not often
used together with CSS anyway like scoreboards
or display panels.

Therefore I think, there is a need of absolute units
in CSS for a lot of applications. The already
existing units, especially centimeter and millimeter
should not be obfuscated with a second definition
related to reference pixel. The used of absolute
units should not be frustrated with contradictory
definitions and implementors should be helped to
implement them properly for devices like screen,
where this is possible. However there are
applications for device pixel and user-coordinates
as well, resulting in some more units to be
defined, one for reference pixel, one for device
pixel one for user-coordinates.


Olaf
Received on Wednesday, 15 December 2010 17:31:01 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:35 GMT