| Abstract: |
In the Amazon Kindle e-books, readers are automatically directed to a linked
e-dictionary which enables the reader to obtain the meaning of the word without
having to leave the book being read - the first three lines of the dictionary article are presented to the reader on the same screen as the open book. If this information is not sufficient to enable the reader to obtain the correct information to satisfy his/her information need, (s)he can follow a link to the full dictionary article. This is an excellent feature to help readers obtain relevant information in a text reception situation.
It works perfectly in most cases. However, unfortunately there are many cases where
the reader is not presented with relevant information or even presented with incorrect information, based on incorrect part of speech identification, linking to the incorrect lemma or a number of other categories of incorrect linking which are discussed in the article. Following a categorisation of such problems, we provide a number of suggestions for possible solutions to the problem. The solutions include the provision of tagged corpora, on-the-fly tagging, a menu of items to enable the reader to disambiguate between various options, etc. Each of these options is discussed in terms of what the option entails, technical issues involved in the specific option and the feasibility of the option. The article concludes that not one option will necessarily resolve all problems and that a multi-pronged approach may be necessary to optimise the automated linking of words in the Kindle e-books to the e-dictionaries that are provided. |