XML from complex and simple layouts

[Rowan_Ferguson@adobeforums.com]

From my experience, it seems that a complex layout document (many text frames, images etc) needs to be manually tagged from top to bottom to get a satisfactory XML file.

On the other hand, a fairly straight forward document that is only 1 connected text frame (perhaps a scattering of images throughout) might be more suited to the "Map styles to Tags" automated method for generating structure.

So based on this, a small complex newsletter for example, would be best tagged manually while a much longer document (like a thesis) might handle the automated method.

I'm just after opinions on this, and what others have experienced?

[pjredman@adobeforums.com]

Rowen,

We've had good success tagging small complex documents (fact sheets) with XML, although it does require some patience. This is not an entirely automatic process but with a combination of the built-in XML tools, some simple scripts, and a little bit of hand tweaking I can tag a fairly complex 4-page fact sheet in about 5 to 10 minutes.

Phil Redman

[Rowan_Ferguson@adobeforums.com]

Hi Phil,

I'm interested, what "in general terms" do you achieve with scripting that helps output XML?

[pjredman@adobeforums.com]

Rowen,

We use scripts to set and execute the style to tag mapper, add attributes to XML elements, such as Alt attributes for Figure elements, and remove unwanted elements, like running heads, from the document structure. This could all be done manually but the scripts are much faster and more reliable than a human operator.

Phil

[Rowan_Ferguson@adobeforums.com]

I see how that could be handy for wrapping container tags around blocks of content, something that doesn't seem easy to do the human way.

Have you had much success tagging tables?

[Jacob_Stevens@adobeforums.com]

Rowan,

Yes, I broadly agree: I work with long articles, which have a reasonably complex structure but always come in two stories (main body of article and footnotes). Tagging by hand wouldn't be feasible, so I map styles to tags. There are currently come problems with ID crashing when mapping character styles to tags (discussed in a thread called 'XML Export'.

When it works, the output I get is well-formed but not valid, as InDesign does not yet offer enough control for the automated tagging. One solution is to use scripting, as Phil does; the solution I use is post-processing the XML output with an XSLT.

With a range of different stories on one page, I would have thought that manual tagging would be the only option.

Jacob.