Steve Pugh said:
1 pt = 1/72 of an inch.
So text sized in pts can not work on both a small palmtop screen or on
large wall mounted screen.
In practice the various browsers and operatrng systems map pt to
screen pixels
Not all, which is why there is an apparent difference in the
rendering of "pt" between, say, Mac OS, GNU/Linux
(browser-dependent) and Windows.
Well at the moment they all map 1 px to 1 screen pixel. Which only
gives consistent results when compared to things that are inherently
sized in pixels, such as images. The actual size of a pixel varies
widely from user to user and with high resolution devices coming into
use text that is readable on conventional displays may be vanishingly
small on the new displays.
And of course when one reads the CSS spec one discovers that 1px is
not supposed to be 1 screen pixel, but is instead supposed to scale
according to certain factors. But browsers don't get that right at the
moment, if they ever start getting it right who knows what problems
will crop up with old designs where px sizing was assumed to be fixed
to screen pixels.
My reading of 4.3.2 is that the rescaling generally applies to
cases where a typical computer screen is *not* the device being
used to view the results of the user agent, so mapping 1px to 1
screen pixel would seem very appropriate--though I'd agree that
given a very high (or very low) resolution screen, a browser
ought to (but doesn't) rescale. Still, they at least partially
get around this by generally providing manual rescaling
capabilities (maybe to font-sizes only, though; unsure).
Who ever said that the pt was preferred over px?
No one; my misunderstanding/misreading, apparently.
The better units brucie was referring to were em or %.
Is % really ever a good choice for font-sizes? Are there
situations where it would be ideal to have your font-size scale
automatically with every window resizing (assuming its not within
a fixed-width box of some sort)?
Yes, the user's chosen default size is a good starting point.
Use Google Groups to find the millions of threads on this topic that
have already taken place.
Mm. I've made a mistake (no nead to read the google groups; your
arguments combined with a more careful reading of the spec in
sections which I must have forgotten) have convinced me. I have
often found it convenient to approach HTML from a typographer's
point-of-view, as it seems that what is right in one is often
right for the other; though this would definitely be an
exception.
-Micah