On 27.10.2006 09:09 Dave Pawson wrote > On 27/10/06, Andreas M. <[EMAIL PROTECTED]> wrote:
> If you assume the frame is a wrapper for the image, then could you use > that information to set css properties on the image, rather than > assume it is structural markup? Thanks for the advice, but it seems, this won't help me. The frame is not a wrapper. It can hold text below the image, like: +-------------------+ | +---------------+ | | | | | | | | | | | | | | | | | | | | | | | | | | +---------------+ | | Figure 1: blah | +-------------------+ and OpenOffice (for example) handles it as structural element, auto-numbering "Figure N". I would like to know some background on why it is being implemented the way it is now. Anyone from the folks involved in the background maybe could clear things up ? > The other way would be to only action the frame if not within a para tag. > Two templates, one for p/frame another for frame without p ancestor. I think I know what you mean. But this is inconsistant. There is no way to make sure, the author did not want to have a frame at that position. What if there is para-text inside of it and a picture (like shown above), right within a frame. This would mean a construct like this (which is invalid): <span><img/><p>Figure 1: bla</p></span>. And I would need to transform this, since it is valid content. >> Also, I see no way as how to deliver the 'width' and 'height' >> attributes to >> the <img>, [...] > So don't, if it doesn't make sense. > Let the image be its natural size. It is a common habit to tell the web-browser the real size of the image so it can use these values for correctly laying out the text, even before the image has been received, without forcing the user to find the passage again, after the images were downloaded and the text re-layouted (jumping text) -- Bye, Andreas M. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
