Default Fonts in KDE

[Friedrich W. H. Kossebau]

Am Sonntag, 17. Juli 2005 03:53, schrieb Germain Garand:
> Le Vendredi 15 Juillet 2005 14:18, Allan Sandfeld Jensen a écrit :
> > On Friday 15 July 2005 01:14, Dominik Huber wrote:
> > >  > 104341 and 151816, but Gnome has no voting like KDE-bugzilla. But if
> > >  > somebody here already has a gnome-bugzilla account he could change
> > >  > at least the gnome-version to a current one and raise the severity.
> > >  >
> > >  > Andreas
> > >
> > > Owen Taylor has reacted:
> > > http://bugs.gnome.org/show_bug.cgi?id=104341#c16
>
> [...]
>
> > On the other hand, I tend to agree with him. DPI as a physical value is
> > problematic and often undesirable.
>
> But DPI *is* a physical value and nothing else. We cannot seriously go
> around and reinterpret it "as" any other random concept of our liking...
>
> When asked for 4 square centimeters of screen real estate, there's no
> "distance from screen", or other dubious "angular size" as invoked in the
> linked bug report... it means "please, give me 4 square centimeters of the
> plane of current screen, thank you very much".

Well, as long as we mix pixel sizes (done for a lot of widget styles, in html 
etc.) and real world measures (like font sizes) we do not really work with 
physical sizes. Problem is that the screen rendering is still bound to match 
full pixels to look good due to screens' low dpis. So pixel numbers and font 
sizes make up a strange pair.

> The only sensible way to achieve that is to obbey the system settings, not
> to invent our poor own!

At least, if they are poor.

> > I believe we in KDE should allow a
> > dynamic DPI such as in Windows to let users select their text size level
>
> but this surely can be implemented as a zoom factor, i.e. unrelated to
> DPI...

Seconded. Or as a new measure "Perceived dots per inch" (calculated by real 
dpi, distance of the viewer and whatever suited, e.g. for beamers). But 
making it simpler as it is is and stays, uhm, suboptimal. 
Hiding some system's complexity is okay, but preventing it is something 
simpler and inventing some bad abstraction which breaks once looked at a 
little bit more should never ever be done. It makes things worse in the long 
run. 

The X setting for a display's dpi should be correct. If one needs a further 
measure he should add one instead of redefining another, others depend on. 
That one is a very dirty hack.

Friedrich
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20050717/35a6236b/attachment.sig>