What is HARVEST ?

    HARVEST is an open software package, first created by Colorado University.

    HARVEST is a so-called decentralized search engine. This means, that it does not index the data from one single place, but it can be installed on different servers, where it carries out a full text inquiry.

    Why do we need a decentralized search engine?

    The 'classical' search engines (Google, Altavista, ...) are mostly global and not specialized.
    A series of fundamental problems result from this:

    1. Many "hits" don't really hit, the user to be threatened to be drowned in the flood of answers.
    2. Many database entries are outdated totally, the links don't work anymore.
    3. The search engine supports the user only little or not at all formulating search inquiries since no know-how can be put into the search engine.

    A specialized machine which only searches in specific fields determined first is favored in this case. Subsequent the locally gained data can be exchanged between the different Harvest systems, Harvest uses the so-called SOIFormat for this.

    The advantage of this procedure is clear:
    Because the indexing is carried out locally, no unnecessary traffic slows down the net. The subsequent exchange occurs with packed data; the net is burdened as little as possible.