In contrast to the ClassicalModelOfInformationRetrieval, the berrypicking model asssumes that as users search a document collection (possibly a number of different searches with different queries), they collect and save useful tidbits (both useful references and actual information). The important difference is that the goal is not to generate a final set of retrieved documents, but rather to collect the “berries” of information gleaned along the way.
In the paper1 that introduces berrypicking, Marcia Bates also introduces the concept of an evolving search. In the ClassicalModelOfInformationRetrieval, users have an unchanging “information need” which they are trying to express as a query. By contrast, in an evolving search, users might execute a query, skim some of the resulting documents, and actually learn things which slightly changes their information need. They then refine their query, not only in order to better express their information need, but also because the information need itself has now changed.

In this picture, you see that the “berries” of “documents, information” are gleaned at various times within the larger series of searches. In contrast, in the ClassicalModelOfInformationRetrieval, one would expect that the user would simply be refining their search until, at the end of the process, the set of documents returned by the last query contains everything that the user “keeps”.
To summarize, there are two differences between this and the ClassicalModelOfInformationRetrieval:
see also: UnvoicedConnection.
Quite often my search terms change because: I’ve learned from some of the earlier documents some of the specialized terminology in a field.
(example) For example, I wonder what sorts of map projections (Wiki:GeographyExample , Wikipedia:Map_projection) could be printed on a nice flat piece of paper and then folded into something vaguely sphere-like. Or in other words, what sorts of vaguely sphere-like shapes can be unfolded into a flat surface.
I’ve learned many terms related to these kinds of shapes over the years:
(Wikipedia:Talk:Polyhedron shows that I suspect there are a few more terms that, if I only knew them, would help me search for maps that use those kinds of shapes).
(/example)
There are many other reasons search terms change (an “evolving search”). I think it would be nice to describe more of them here.
Berry picking lends itself to a nice analogy with 2 opposing problems in searches:
analogous to:
This gives me an idea: a SearchJournal: Something you can use to document the process of a search. What you were thinking, where you went, what you found, how it changed your search.
Footnotes:
1. Marcia Bates. THE DESIGN OF BROWSING AND BERRYPICKING TECHNIQUES FOR THE ONLINE SEARCH INTERFACE. http://www.gseis.ucla.edu/faculty/bates/berrypicking.html