Lists Home |
Date Index |
It is not useful to pit the 'common' user against the 'guru'.
It is useful to compare the approaches to deriving and applying
1. A bottom up approach is effective in the same way having
locusts harvest crops is effective: self-directed, large
numbers, slightly messy but fast. A single harvester grabbing
single rows is more expensive but may be doing other tasks along
2. An approach like Google Base where the initial inputs are
examined then guided is a mix of the two approaches and likely
to be the most effective overall as an ontological framework
exists and mapped but the common user doesn't have to learn it.
3. The difference is a priori learning vs just-in-time commitment
to the ontology. A guru designs an ontology, someone designs and
implements a mapping strategy, and the user is given a choice of
ad hoc inputs and selections. This is a potentially noisy approach
but over time, it is possible to derive the statistical deviations
and assign confidence values.
4. Use determines application. For some large percentage of
cases, the mapping noise isn't significant if the application is
search unless the return value is used for a significant process,
the outcome of which is critical to some other process.
To understand if and where the semantic web adds value, you must
know the value proposition of the process(s) relying on its
service(s). This is where the semantic web meets service oriented
The common quality of webs of any kind are holes.