This list of technical writing myths has a decidedly DITA slant; I don’t necessarily like the idea of DITA driving what is and isn’t acceptable practice for technical writing. That being said, I endorse the information provided about myth #4:
4. The Callouts on Graphics Myth
If you want to reuse the same graphic in multiple publications and even multiple languages, it is a good practice not to put callouts in the source file of the graphic itself. Instead, you place your callouts “on top of” your graphic in your text editor (Word) or DTP program (InDesign, FrameMaker, …). This is not supported in DITA. Therefore, if you need to use callouts in graphics, try to use language-independent ones (A, B, C…) in the source file of the graphic and put the explanation of these callouts in a table below the graphic in your DITA XML content. [Yves Barbion]
It’s not just the DITA standard that doesn’t support callouts placed with a document processor’s drawing tools. We have a client for whom we created a custom XML structure for FrameMaker content, and that content contains many graphics with callouts. The client translates the content into multiple languages.
In the previous workflow in unstructured FrameMaker, the client placed the callouts in the anchored frame with FrameMaker’s drawing tools. However, if you do that in a structured FrameMaker workflow and save content out to XML, FrameMaker by default saves each image with added callouts into a new CGM graphic file that no longer has editable callouts. There can’t be complete round-tripping between XML and FrameMaker because the graphics aren’t preserved in this scenario.
Therefore, the rule for this client is that an anchored frame can contain just one image file imported by reference. Period. No other text, text frames, or anything. If callouts are necessary, they are included as part of the source graphic; the client adds numbered callouts in Illustrator. A numbered list specifically for explaining those callouts follows the graphic.
The XML generated from FrameMaker points to the image files, and FrameMaker doesn’t need to generate new graphic files to include any callouts added with FrameMaker’s drawing tools. (If the directory containing XML files includes a CGM file that FrameMaker created during export, that usually means there is an anchored frame somewhere that contains something more than just a referenced image file.) Another huge plus: because the callout explanations are part of the text, they are translated without changing the original graphic.
Separating callouts from your images and making them part of the text is smart for any workflow because of localization issues, and it pretty much becomes mandatory when you’re establishing an efficient structured authoring environment (and not just those based on DITA).
P.S. One other thing to think about with graphics in structured authoring environments: if you use XSL to transform your XML into online formats, determine if web-ready graphics (such as PNG files) will work in your print/PDF and online content. If so, you eliminate the need to convert images to formats for online viewing.