Currently I am reading here to learn more about XPath. I need this knowledge to be able to navigate through PLS and SSML documents. To be more precise: at the moment, I don’t use PLS and SSML. I am just using XML. And why is that? The answer is easy to understand: XML is less complicated than PLS or SSML. I am trying to steal the tags that are part of the PLS and SSML specifications. And I am trying to implement those tags into XML. In the long term, I would like to employ PLS and SSML directly. But step by step. Before I do that, I have to know more about XSLT and XPath. This is the basic knowledge.
To be a bit more specific about PLS: I want to transform a pronunciation dictionary (UTF-8, PLS, IPA) into a different document (UTF-8 or ASCII, PLS and HTK compatible, SAMPA like). So in the end, I would have two or three different versions of the same pronunciation dictionary. And to achieve this goal, I need XPath. I have to navigate not only through the elements and the attributes of the elements but also through the IPA-symbols themselves. And the details are pretty complicated, I assume. But I have found an example (I set a hyperlink to this example in a different weblog entry) that is very promising. There is a special XPath syntax that should help me to find the solution. And I am confident that XPath is compatible with UTF-8. So there shouldn’t occur any problems at this layer.