FOA Home
How could we possibly know so much, about the story or argument an
author is attempting to get across?! Or about just what question a user
has in mind, and why they want the answer?! In fact, more and more
evidence is making itself available to make just such determinations.
Virtually every author now composes using a word processor, and even if
they are only interested in ultimately producing a paper
document, it generally carries with it much of the intra- and
inter-document structure discussed in Chapter 6. Further, it may be
possible to infer even more of their thinking if the process of
word-processing is allowed to leave a trace with the document.
Similarly, browsers retain more and more state all the time. BOOKMARK
LISTS and SEARCH HISTORY files are the most obvious examples.
Less obvious to most users are the COOKIES left by WWW servers on
the user's client, so as to better identify just when they come to
visit.
The point is that there is actually fairly rich streams of
information available about both author and user, above and beyond the
query and document text with which we have been primarily concerned in
FOA. Of course there are some computations over these two data sets that
can be done a priori, before any search commences. But the more
interesting computations are those that can only be done on the occasion
of the particular query. A sketch of this perspective is shown in Figure
(figure) with the user and their active seeking behaviors
creating a PORTAL through which corpora can interact.
Here we
propose that in fact each user's query, each bit of RelFbk and move
from one document to another effectively creates the
opportunity to juxtapose two corpora in ways that, until that
moment, had never been analyzed in just that way before.
Top of Page
User as portal