To understand the purpose of this thesis, it is important to understand the information we are working with. While we call these objects simply documents, the term is loaded with meaning. This chapter, among other things, will discuss the philosophical (as well as more reality-grounded) aspects of documents [7].
There has been much research of the nature of documents, and an all encompassing study of these definitions is well beyond the scope of this thesis. For our purposes we can condense a number of definitions to give documents following definition: ``A source of information in any form which is archivable.'' This broad definition applies, therefore, to all media formats as long as they maintain some sort of persistence. It is notable, however, that Haystack functions best on documents with some text form.
With this definition of documents it is now possible to explore their internal nature. We make the additional argument that documents are composed of stateful properties and relations. We can then leverage this state to represent documents within Haystack in a manner that allows for improved searches.