Re: Sizing

On Sun, Sep 17, 2006 at 04:44:43PM +0200, Axel Dahmen wrote:

> >      -- http://www.w3.org/TR/CSS2/syndata.html#value-def-length
 
> Right. But instead of using user agent magic to rescale, user agents should
> use CSS when rendering, making it possible to override/change these values,
> e.g. by DHTML.

Why? The page author can't know what values are suitable for the
display device.
 
> > This seems to be moving responsibility for defining the DPI of the
> > system from the system administrator (who has a chance of configuring
> > it correctly) to the document author (who, in the case of the WWW at
> > least, hasn't a hope of doing so).
> 
> AFAIK, currenty no system administrator has any chance of configuring
> anything regarding scaling. At least not in current browsers.

In Firefox/Linux (for example): Edit > Preferences > Content > Fonts &
Colors Advanced > Display resolution has some influence.

More usefully, I can configure my X server to specify the physical
dimensions of the display (with DisplaySize).

So the facility is there, although I don't know how much advantage
browsers take of it.

> By introducing the "scaling" property, he has. He could add device
> dependant scaling to the user (agent) style sheet
> (http://www.w3.org/TR/CSS21/cascade.html#cascading-order).

I'm instantly dubious of anything which is only of use in a user
stylesheet. 

What if the user has multiple printers, or a printer which supports
multiple output resolutions? They would have to change their
stylesheet every time they changed print or resolution. Far better
that the client just read the data from the device.

Additionally, being about to accurately represent absolute lengths on
the display is not a problem limited to webpages. Why create a web
browser specific way to provide this information when it is needed for
other applications? (See the level of support among webservers for
<meta http-equiv> in HTML).

-- 
David Dorward                                      http://dorward.me.uk

Received on Sunday, 17 September 2006 15:34:40 UTC