-
Notifications
You must be signed in to change notification settings - Fork 0
/
notes
26 lines (16 loc) · 871 Bytes
/
notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Obviously this requires a focused crawler.
Now we work on the important business of identifying redundant text,
and orienting the search.
The search I feel it in my heart must derive in part from the
questions and statements the user has been making. In other words,
there are existing contexts, unresolved questions, open areas to
investigate, and so forth.
The goal of the search interface is to help the user to read that.
And, in other words, to always keep the reader busy.
Textual quality metrics like LSI should be considered.
In lieu of a really advanced telepathic system. The goal is to get
some kind of text-streaming system working, tonight. For it is from
reading that lots of good will come.
Textual redundancy metrics.
Automatically build a list of what the user is interested in reading
based on their writings...