Re: CSS3 Web Fonts issue with ?block on downl oad?

On Thu, May 7, 2009 at 11:59 AM, David Hyatt <hyatt@apple.com> wrote:

> I never said the solution couldn't be improved upon, and I never opposed
> the suggestion of progressive streaming.  In fact I think that's a good
> idea.
>
> With existing technology, the choice is between always flickering all of
> the text on the page, or having text pop in and risking some visual jumping.
>  I never said this sort of jumping would never occur in the latter case, but
> the choices are between "always look bad" and "occasionally look bad" when
> dealing with what we have currently.  WebKit has opted for "occasionally
> look bad."
>
> An even better solution is obviously "never look bad"
>

David,

What do you think about the idea of having some basic metrics load first
(somehow), and using that on a synthetic font, so that a person could begin
reading and then continue with a non-jarring change once the rest of the
font data loaded? I guess in an EOF, that would just be the first part of
the font to load, but in regular fonts it would need to be either a separate
file or embedded in the CSS?

You wouldn't have to support the whole synthetic font specs, just enough to
get the glyphs in the right positions without overlapping.