[Melanie] In my opinion eLearning is indeed really fast and you will get the most up-to-date information, but when you are not really motivated it can be really difficult to stick to your material.
[Swathi] I agree with this point. The way in which material is studied depends very much on the user’s interests and motivation (although earning money is for some a rather good motivator).
[Melanie] In the article about eLearning, they name all but the trust layer of the Semantic Web layer cake as being important for eLearning. I think also the trust layer is very important, because you really want to know if you can trust the contents of some source or not.
[Melanie] To me, the article about SWAN was not really clear. I understand their goal (finding a cure for Alzheimer Disease), but why would they need the Semantic Web and ontologies?
[Arlette] So, the Semantic Web claims to provide general solutions to semantic interoperability. However, because conversion and alignment efforts depend highly upon resources plus technical and fucntional requirements of the applications, we cannot really speak of 'general' solutions. A standard model will for example never be able to capture particular semantics of a vocabulary.
[Arlette] Current ontology mappers like Falcon and S-Match, produce poor results in real applicatons. This is partly caused by differences in depth levels of different thesauri and by the loss of semantic information in the formalization step. How can we solve this problem in the future applications?
[Matthijs] Sonntag et al: “Humans tend to keep their contributions as short and efficient as possible”: do humans want to communicate in natural language to computers at all?
[Matthijs] In ontology mapping, a score of only 46% correct matches is already enough to conclude ontology mapping is possible
[Matthijs] The Semantic Web turns out to be useless when building real applications.
[Arno] National and international standards should be developed and enforced for semantic (web) applications for several important knowledge domains like medicine, law and higher education.
[Arno] We should focus more on integration of SW applications and information collections, instead of building separate applications for each, which only adds opaqueness to the semantic web.
[Arno] A big obstacle for the development of the Semantic Web is that computers are not sufficiently able to convert speech to text and comprehend semantics from text in general; future SW applications will often be controlled by unstructured input (speech/text).
[Swathi] The article by Stojanovic does explain that eLearning must be fast, just-in-time and relevant for businesses, but is that true also when it comes to eLearning in education? I think there’s significant difference in eLearning in these two domains, which should not be forgotten.