Updating the local index is not hard; a "robot" like program can be initialised to traverse
the filesystem looking for documents to index,
recording locations and keywords.
There are a few ways of extracting keywords from a document. One way is by extracting pre-defined keywords,
another is to parse the document through a language-recognising program to extract the keywords.
Pre-defined keywords mean that the keywords are already defined and placed within the document in some standard format. This way, the keywords would describe the document better. On the other hand, keywords extracted by programs might not accurately describe the document [12]. For example, a program might extract the keyword "heart-attack" from both sentences:
Heart-attack is caused by narrowed arteries, and it is killing more people everyday ... ... As she turned him around onto his back, she was sure that the heart-attack had killed him painfully ...
Although the first sentence is about the medical condition, the second is a sentence from a novel, and has
nothing to do with heart attack as such. Despite this disadvantage, keyword extracting programs
are the next best thing for documents that do not have pre-defined keywords, because having some keywords
to index a document by is better than nothing.