Authoring and Annotation of Web Pages in CREAM .pdf

Siegfried Handschuh for building such a knowledge portal. The potential users provide knowledge, e.g. by annotating their web pages in a decentralized manner. The knowledge is collected at the portal by crawling and presented in a variety of ways. Similarly, our second scenario here, we have used semantic mark- up in order to provide knowledge about the TIME (telecommuni- cation, IT, multimedia, e-business) markets in the TIME2Research portal [33]. Thus, we have created a knowledge portal for busi- ness analysts. The principal idea is that business analyst review news tickers, business plans and business reports. A considerable part of their work requires the comparison and aggregation of sim- ilar or related data, which may be done by semantic queries like ?Which companies provide B2B solutions, when the knowledge is semantically available. At the Time2Research portal they handle different types of documents, annotate them and, thus, feed back into the portal to which they may ask questions. At the start of these two case studies, we had the intuition that these cases may be easily supported by a simple annotation tool. During the course of the projects, however, we found that we had to face many principal problems. Also, we found that the two sce- narios would never succeed in the end if we did not give people the possibility to easily create documents and their metadata in one step. The principal requirements for a-posteriori annotation as well as for the integration of web page authoring with metadata creation can be outlined as follows:  Consistency: Semantic structures should adhere to a given ontology in order to allow for better sharing of knowledge. For example, it should be avoided that people use an at- tribute, where the ontology requires a concept instance.  Proper Reference: Identiers of instances, e.g. of persons, institutes or companies, should be unique. For instance, in