Re: Propsed length unit: angles as visual angle

Andrew n marshall wrote, at 9:09 -0800 on 21.11.97:
> Style sheets do not have any length that correspond to how a page my be
> viewed.  The relative values that do exist relate to some preexisting
> absolute value.  For example, %, em, and ex correspond to the inherited
> font size.

But, as you observe below, this inherited font size needn't be supplied by
the author in any fixed unit, but can be relative to "medium", or simply
"1em", whose value is chosen by the user (or the blimp pilot) as a
comfortable norm. I try to illustrate this point (clumsily perhaps), on
this page: .

> Ideally this implies every display device should have knowledge about how
> it is viewed, and since it would be difficult to continuously update
> accurate measurements, it should only be assumed it is an average value.
>  While this isn't immediately available on most systems now, a good guess
> can be derived from the default font size for a device, making it not
> unreasonable to implement.

Unfortunately, I see little systematic attention being given to such issues
by the leading UA developers. Note that Netscape's 120-ppi "virtual pixel"
flouts the W3C recommended 90ppi value, and that Microsoft now uses
*7-point* Verdana on their corporate site, which is impossibly tiny on a
Macintosh, where 1 point=1 pixel (this is in IE3 for Mac, of course, where
I have no means short of canning CSS to override this specification).

Todd Fahrner

Received on Friday, 21 November 1997 19:59:13 UTC