So there exists a problem in the last 4 places that I've worked. I've implemented Maven as a build tool at all the places, and eventually people come around to seeing what it provides. But the documentation issue is still a pain. I primarily use Subversion for SCM, and committing binaries to it in the form of the ubiquitous M$ Word document is painful, both from a philosophical standpoint and from the occasionally checkout delay when a document undergoes a lot of revision. Developers being the lazy sorts that we tend to be aren't interested in writing what little documentation they actually provide in anything other than the corporate mandated word document. As for anyone other than developers, forget it. Many of the BAs in my company don't even understand that other document formats even exist.
In the end, for the several hundred maven projects that we have here, all the real documentation is in word docs committed to the subversion tree that then have to be bundled into src/site/resources and then links made to them in the site descriptor. Since the generated site is one the most useful elements of a maven build (at least for us), the documentation being part of the site is preferable to having a download link for a piece of documentation. So my question is does anyone has a template or transformation tool for taking Word documents and turning them into clean XDOC? I just spent about an hour using Open Office to export a word doc to XHTML and try to clean it up. The results are acceptable if unspectacular, but that's more likely due to the template that the original docs were written using. I'm not particularly familiar with Open Office's tools other than the standard ones, but I know that people regularly produce filters and such. Has anyone here written a filter/export tool/XSL transformer/anything that could automate this task? Any thoughts on this? Thanks! Mykel
