Traditional Search Engines have several general problems as they
do not take the meaning of the data into account. If semantic data is given in the Web searching will be more efficient and precise than ever before as Knowledge Representation can be used including Ontology and Linked Data. If a Search Engine is queried with ambiguous terms it shall offer to refine the Query.
Further problems arise as the Query String need be somehow interpreted and the context of the user need to be taken into account, i.e. her/his interests, what has s/he ask just before or what is her/his usual preference. The entities that are connected to the terms of the Natural Language query string, they have to be correctly identified. This mapping between terms and entity identity is difficulty and usually error prone. What is need to solve this is some kind of automatic disambiguation. This process has to be done in an automated way so that the user does not notice it or on the other hand the user has to be asked without interfering with the usability of this process. Moreover, the personal Context need to be taken into consideration.
Thus, the general problems of document retrieval with conventional Search Engines can be classified into the categories: