RE: font sizes in ems: a clarification

I see the problem in IE4.  I'll attempt to fix the fact that a
non-"medium" default size is not being used to calculate font sizes set
in EMs.

BTW, my email address is "cwilso@microsoft.com".  Please don't send mail
for me to "cwilsoN" - Carolyn Wilson doesn't really need a copy.  (Sorry
again, Carolyn.)

	-Chris
Chris Wilson
cwilso@microsoft.com
***

> -----Original Message-----
> From:	Gayle Kidder [SMTP:reddik@sandiego.com]
> Sent:	Friday, July 18, 1997 12:49 PM
> To:	Chris Wilson (PSD)
> Cc:	Carolyn Wilson; www-style@w3.org
> Subject:	Re: font sizes in ems: a clarification
> 
> Chris Wilson (PSD) wrote:
> > 
> > Gayle, I don't understand what the problem you're seeing with IE4.0
> pp2
> > is. 
> 
> This discussion has drifted a bit from my original intent (as things
> will do), so let me get back to my original problem:
> The original page I posted had no BODY default size on purpose. The
> idea
> was to force the browser to use the reader's default font size. This
> neither ver4 browser appears to do at the moment. I tested this by
> trying various fonts on the page and choosing to give them first an
> absolute size of 10pt and then a relative size of 1em. My
> understanding
> from the group discussion was that the relative size should be
> relative
> to my user default. I can change my default size from now til
> Christmas
> but it has no effect on the rendering of the lines spec'ed as 1em. So,
> either:
> 
> 1) The idea that setting an initial font size in ems will respect the
> user's default size and scale font sizes accordingly is fallacious, or
> 
> 2) The browser makers have not seen fit (or clearly enough) to
> implement
> it.
> 
> (The third option, of course, is that I misunderstand the
> specification.) I suspect the browsers are either using a browser
> default setting rather than my reader's choice, or the size is
> relative
> to the previous element rather than the parent element. In either
> case,
> specing in ems appears useless.
> 
> There is the separate problem, discovered along the way (and
> inadvertently complicated by the lack of inheritance in tables) of
> whether an em size on any particular font renders the same as a point
> size. On this,
> 1) MSIE appears to render an em the same size as a point. The glitch I
> captured from David Perrell's demo, however, demonstrated that if you
> change font size using the View/Fonts menu, it only affects the first
> rendering (which was font-size: 1em) but does not scale the rest of
> the
> page accordingly (i.e., font-size: 2em is relative to the original
> font
> size but not the new resize). You'll have to revisit his demo to see
> this.
> 
> 2) NS4 does not render an em the same size as a point, nor does it
> seem
> to use the reader default. This led to my assumption that it was
> scaling
> the font larger for an em and therefore one needed a scale-down factor
> to spec in ems. This latter you can see with NS4 on a new, cleaned up
> demo at:
> http://www.beachmedia.com/www/emdemo1.html
> 
> I hope this helps clarify the problem as I see it.
> 
> Gayle Kidder
> 
> 
> > As far as I can tell, we render this correctly - in fact, we render
> > your first page correctly if you set the font-size explicitly on
> <BODY>
> > _AND_ <TD>.  The first line (explicit 10pt) is the same size as the
> > second (1em), as I believe it should be - and the third line (.9em,
> > etc.) is smaller, as it should be.  These do not scale when you
> select a
> > different default font size (I don't believe they should).  Am I
> missing
> > something?
> >

Received on Friday, 18 July 1997 17:00:25 UTC