When images are part of a document, effort must be taken so the result looks good in the converted document. Often, more than one step is necessary to bring an image from a source to the target document, usually GIF and/or JPEG when heading for HTML. A good set of image manipulation tools will help in making this step easier. But still, one has to know a little bit on what's important when generating images.
First and most important is, that every detail on the picture is visible. Ways to archive this are switching on anti-aliasing if the tools and image formats support it. Anti-aliasing has a positive effect not only on fonts and their readability, but also lines, circles and other graphical objects look less crude, whether they are color or black and white. See Figure 7 for an example.
When a image contains text, the notes on fontsize explained in the previous section apply.
Furthermore, when an image consists of line drawing, the width of these lines shouldn't be taken too narrow or to wide, to keep details and not stroke them out by too thick lines. This is usually no problem when going for PDF, as this format contains good mechanisms for sizing images itself, as well as anti-aliasing is usually done automatically by the document viewer.
But when converting into a pixel format such as GIF and/or JPEG for use in an HTML document, keeping details may be a problem because to have all the details visible on the picture, the dimension of the image may exceed the size of what can be viewed even on a 640x480 display. To solve this problem, it's easiest to put a small (thumbnail) image on the page itself, and then create a link to the image with a higher resolution. That way, everyone can watch the details in his own demand, and no time is wasted to download some big image the user doesn't want to watch.
The same kind of time-saving can be archived (in addition) by reducing the number of colors an image uses. It's of little use to insert any true- and highcolor-images with millions of colors, when most of the users have displays which hardly support 256 different colors, and even these colors are shared by several images, documents and/or applications.
Beside all those visible aspects of the contents of an image, there are also some guidelines to regard on inserting them into documents.
Most important, every image should have some text outlining what it's good for, either directly above of below it. (Of course it should be only above or below throughout the whole book, not mixed!) When converting into HTML, one might also set an additional ALT-parameter to the IMG-tag (<IMG ALT="sometext" ...>) to give a textual description of the picture in case the browser used to read the document has images disabled or doesn't support them at all.