Re: style sheet scalability (was: Specifying style notation in <link>)

lilley (lilley@afs.mcc.ac.uk)
Tue, 5 Dec 1995 18:15:26 +0000 (GMT)


David Seibert said:
> On Tue, 5 Dec 1995, lilley wrote:

> > Further, a well constructed antialiased image (at, say, 72dpi) containing
> > text will be more legible if displayed 1:1 on a 110 dpi screen - thus making
> > it too small - than if it were scaled by a factor of 110/72 by pixel
> > replication - which would make it bigger, but drop the quality right down.

> Legibility should normally be increased with a larger display area,
> assuming you don't get make the image too big for your display ;), as
> long as the number of pixels you use on your screen is an integral
> multiple of the number in the bitmap.

As I showed earlier in this thread, that depends on your scaling algorithm.
For pixel replication then sure, integer multiples are a way to preserve
most of the quality/ legibility while making the image bigger. Which is
why I said:

> > Remember
> > we are talking about scaling factors of at most 50 - 200% and more
> > likely 80 - 125% so Walter's suggested hints of integer scaling do
> > not apply.

> If the author specifies a minimum
> display size for an image, that should not be a problem to honour; if the
> granularity of the image is higher than the granularity of your screen,
> you could expand the image (probably by order 20% or so) to cover the
> same number of pixels,

The act of doing which would reduce the legibility of the image and (in
particular) any text in the image id the faster and more common (I would
say even universal) scaling methods are employed.

People are saying that bigger is clearer, and I agree *all other factors
being equal*. Trouble is, they are not.

> while if the screen has a higher granularity you
> could remap each pixel to an N x N block of pixels, where N is the minimum
> number to make the image larger than the desired minimum size.

So N=2 in the vast majority of cases. I assume (Fortran!!) you mean
integer N. The only case where N> 2 would be required is an image designed at
or below 60 dpi and displayed at a higher resolution than 120 dpi. I am
hard pressed to think of any system that has display resolutions
outside that range.

> This
> should preserve the minimum size and maintain legibility. As Chris
> pointed out, a fixed size is obviously much harder to render, but it
> would be good to be able to prescribe a minimum.

A fixed (physical) size is much harder to render *well*.

-- 
Chris Lilley, Technical Author and JISC representative to W3C 
+-------------------------------------------------------------------+
|       Manchester and North HPC Training & Education Centre        |
+-------------------------------------------------------------------+
| Computer Graphics Unit,             Email: Chris.Lilley@mcc.ac.uk |
| Manchester Computing Centre,        Voice: +44 161 275 6045       |
| Oxford Road, Manchester, UK.          Fax: +44 161 275 6040       |
| M13 9PL                            BioMOO: ChrisL                 |
| Timezone: UTC        URI: http://info.mcc.ac.uk/CGU/staff/lilley/ | 
+-------------------------------------------------------------------+