[
Lists Home |
Date Index |
Thread Index
]
Something else to consider. I've been reading Tim Bray's
excellent blogs on searching. Sowa makes much of analogical
means to detecting similarities.
1. Can analogical systems be used as adjuncts to search
engines?
2. Given the position Tim seems to be taking that search
engines don't get much bang for the buck out of some approaches
to improving search results (using Google as the example),
does the mean that the Semantic Web doesn't offer much that a
search engine really needs, or is it the case that whereas
a user searching the web gets little benefit, other applications
would? What would these other applications be?
Sowa makes much of the use of analogical reasoning as a means
to create and then select among theories to apply to a situation?
We make much here of "lessons learned" but these are in effect,
"theories we apply" before we attempt to use logical means to
argue our case. So perhaps one application is as a means to
support argumentative debate by having a system that can look
at cases and find precedents. Legal systems could use that.
Another might be analysis of written works (lists and blogs
come to mind) and look for similar positions taken for comparison
to historical results. This becomes the "protect against what
we know that just ain't so" system.
Such a system might require loose searching then benefit by
automated analysis that uses an analogical reasoning engine
to sift and find the most relevant cases, then illustrate
point for point comparison with results summarized.
1. Search for precedent cases.
2. Create theories.
3. Select among the best theories offered.
4. Create the case.
Consider the thread started originally with a poor filtering
engine for finding "bad words" or "bad sites". Such systems
probably use word and site lists, but do not do word sense
analysis or make cases for the classification. If a policy
maker acted on these, they would also have to make the case.
len
|
