The following table illustrates the expected retrieval behaviour for various combinations of attributes. The behaviour with regard to combinations of the truncation and completeness attributes has been especially troublesome in terms of predicatable interoperability. The following table illustrates the expected behaviour.

term = "dog"

use	truncation	completeness	finds
title	none	complete field	titles containing only the single word "dog". Would find "Dog."; would not find "Dogma" or "A dog and bone story"
title	none	incomplete subfield	titles containing the word "dog", but not titles containing words starting "dog…". Would find the previous and "A dog and bone story"; would not find "Dogma"
title	right	complete field	titles containing a single word beginning "dog…". Would find "Dog" and "Dogma"; would not find "A dog and bone story"
title	right	incomplete subfield	titles containing words starting "dog…". Would find the previous and "Dogma and the Christian church."

Additionally, the position attribute determines where in the matching database term the search term is required to be:

use	position	truncation	completeness	finds
title	<irrelevant>	none	complete field	titles containing only the single word "dog". Would find "Dog."; would not find "Dogma."; would not find "Dog and cat"
title	first in field	none	incomplete subfield	titles beginning with the word "dog", but not titles beginning with words starting with "dog…". The title may contain additional words. Would find the previous and also "Dog and cat"; would not find "Dogma."
title	any position in field	none	incomplete subfield	titles containing the word "dog", but not titles containing words starting "dog…". The title may contain additional words. Would find the previous and also "Me and a cat named Dog", would not find "The truth about Katz and dogs"
title	<irrelevant>	right	complete field	titles containing a only single word beginning "dog…". Would find "Dog" and "Dogma."; would not find "Dogma and the Christian church"
title	first in field	right	incomplete subfield	titles beginning with a word starting "dog…". The title may contain additional words. Would find the previous and also "Dogma and the Christian church"
title	any position in field	right	incomplete subfield	titles containing words starting "dog…" anywhere in the field. Would find the all of the previous and also "The truth about Katz and dogs"

The behaviour with regard to other values of truncation can easily be extrapolated. Few servers support "left truncation" or "left and right".

Note that Bib-1 does not address behaviour with respect to initial articles or stopwords. Whether or not term "dog" with truncation=do_not_truncate and completeness=complete_field finds "The dog. or "A dog." in addition to just "Dog." is a local implementation issue. [

Retrieval behaviour with regard to position="first in subfield" and completeness="complete subfield" is undefined.