What I proposed is to develop a support tool to do the model-enhancing work for Growing Books. The difficulty of this project is to retrieve the logical levels and media type information of the Growing Book documents. Naturally, we find there are mainly two ways to grab the information.
The first way, we can use some editor to directly insert tags into documents or do some marking on the documents according to the logical structures and then develop a tool to translate the marking into MAWC tags. -- But it just changes from directly manual operation to another form of manual operation. It's not attractive.
Another way is to automatically adding MAWC tags into HTML documents. This is more convenient for the users, but also more difficult to do because it need to analysis the logical structure and media type information of the HTML documents. My idea is to do it in the following ways. First, we find some patterns of the logical structure of the HTML documents, such as using "1.1","1.1.1" to indicate the level. Second we will translate these patterns to logical structure rules. We get the media type information from the grammar of the document specification language, such as HTML, XML. This kind of information is also being translated into some rules. Then, we can analysis the logical structure and the media type information of the hypermedia document. With the information, to add MAWC tags to the Growing Book documents becomes straightforward. But we have to recognize that there are some risks on the second approach-If the rules only cover the documents written in some specific kind of style(i.e. using , tags to indicate the logical level structure), the tool will fail when it is used to enhance documents written in another type of style. But we can solve this problem by elaborately designing rules to achieve good performance of this tool. I adopt the second approach because it is more powerful than the first approach although there're some risks.