What factors affect the way Wikipublisher processes images in Web pages when these are typeset for print?
Behind the scenes, we go to great lengths so that, in general, images “just work”. When preparing an image for printing, Wikipublisher has to consider a number of inter-related things:
Image processing has 3 stages:
If the image has a left or right float style (or imagesize=sidecaps), we place the caption beside the image, making sure the image takes up no more than 60% of the text width (the Golden Ratio). The image is printed with the caption on the inside margin. Otherwise, for non-floating image styles, captions print below the image. Figures with captions are numbered automatically. If no caption is specified, it uses the image alt text — we assume authors know that is it good practice to specify alt text for all images.
To implement high resolution image substitution, Wikipublisher re-interprets the [[text or image1 -> image2]] link markup. If it finds markup in this form, it uses image2 for the print version of the page. If image1 has alt text, it transfers this to image2.
Several people have asked for a way to fix an image “here” and flow text around it. There is no easy way to do this. By fixing an image in place, we have to trust the author to verify that the image is not, for example, too close to the bottom of the page on which it prints. But the same wiki page can be printed on its own, as part of one or more trails, or as part of a search result. It can also be printed with any of 3 paper sizes. So an author cannot know whether the image will fit where she specifies. It gets worse. If we fix an image “here”, it is possible that an image defined above it will float past it, and the figure numbers will be out of sequence.
1 This is a new feature; by default, images float to the bottom of the current page, the top of the next page, or to a separate float page. (↑)