Details:
Summary:
The publishing sector has changed essentially during the last five years. Earlier, the publishing process was in many cases unique, i.e. a certain content was prepared for a certain publication. In this kind of a process there was not so much room or need for tools that would automate the content manipulation and compilation processes. This has changed. To serve the different channels and to publish personalized content, much more work needs to be done with the same content. Content preparation and compilation should be automated, where possible. This paper discusses two implementations where automation and computer support are created for two content processing tasks. The first case concerns the XML tagging of a news article. The software searches for proper names and acronyms in the text and shows them to the user for classification. The classified words are tagged according to the classification. These tags create a new level of information within the document, and they can be used to improve the searches of the documents or to modify their rendition. The second case shows an application of Kohonen's one-dimensional self-organising map to sort articles in a good sequential order. A logical sequential order is the best solution with devices like eBooks and MP3 players. Some test results are included to demonstrate the performance level of the application. This paper also discusses automatic abstracting. A respective application is currently under development at VTT Information Technology. Automatic abstracting of text documents is very useful in various media applications. Abstracts may be used as part of the metadata, or they can be delivered to the readers as such, for example, to mobile terminals with restricted transfer and display capabilities. It is difficult to fully automate content-related tasks, but the implementations give encouraging results. With automation and computer assistance, content manipulation tasks can be speeded up and the prerequisites can be created for increased publishing through multiple channels and in a personalized way.