Re: SVG 1.2 Comment: 4 Flowing text and graphics

"David Woolley" <david@djwhome.demon.co.uk> wrote in message 
news:200411070914.iA79Eos01911@djwhome.demon.co.uk...
>
>> We want the transformation from semantic to presentational to happen at 
>> the
>> latest possible level, client-side SVG is that level - hence the need for 
>> a
>> known text-wrapping algorithm.
>
> You don't need a known transformation for that.  You only need a known
> transformation if you have already applied the transformation and checked
> it visually at the authoring side.

There is no authoring side, the transformation is done in the client, this 
is so that we can deliver semantic mark-up to the client, however this 
prevents us from pre-computing the rendering.

> If you completely delegate to the
> client, you shouldn't care about the details of the rendering.

That would appear to be an argument for no SVG at all, the reasons for a 
mark-up language for graphics is because it's more efficient on a number of 
different levels, and can have semantic and accessibility information tied 
into it.  We do care about details of the rendering, that's why we're using 
a final form graphics language, but that does require consistency.

Jim. 

Received on Sunday, 7 November 2004 11:52:02 UTC