Erik Hatcher wrote:
Maybe I didn't understand why you suggest to use RDF?
I must be misunderstanding then. The website you linked to had this:
http://www.nongnu.org/esp/#XML_Examples - so I presumed some machine
readable thing going on under the covers.
I've got three different document types: SocialContract, Policy and Imprint.
SocialContract and Imprint are XML formatted content that is machine
readable down to the simple types.
The Policy document type is a document that mostly consists of human
readable text and can be converted to (e.g.) XHTML by a XSLT processor.
The reason for the Policy document not being XHTML formatted is that
it can follow a predefined structure to allow a certain degree of
classification
of its content by analyzing that structure. Also a Policy document allows
for much less arbitrary and superfluous formatting information than an XHTML
document.
You were right in assuming that something machine readable is going on:
A search engine can process imprint, social contract and policies of an
entity advertising these and can consequently allow an interested user
to search for information that was previously unavailable.
You get (hopefully) uniquely identified entities with identifieable
services,
products, contact and address information and, above all, ethical standards.
The ethical standards are broken down sufficiently into machine readable
content that a search engine user can request ethical standards for the
organizations or services he or she is looking for by specifiying groups of
policies or (further policy groups) that are to be implemented collectively
or alternatively, thus allowing arbitrarily complex expressions of AND
and OR clauses for ethical policies and available implementation levels
of these policies.
<http://dict.leo.org/se?lp=ende&p=/Mn4k.&search=superfluous>
You also mentioned the analogy to Java classes and interfaces, which
is something RDF represents as well.
That statement was mostly to correct a wording I used earlier, which
stated that
policy schemes were like Java interfaces and could be implemented. I didn't
think about the confusion this would create for the user, when a policy
scheme
could be implemented by a policy, which then could be extended by another
policy to be, finally, implemented again, by a social contract. That was
a poor
choice of terminology and needed to be corrected.
Creative commons is only related to my project in that creative commons
licenses could be an intersting search criteria. Web services that
offer
content under creative commons could, for example, advertise an ethical
policy that guarantees all content to be available under creative
commons licenses.
What use of the Creative Commons plugin where you suggesting
specifically?
I mentioned CC because it parses embedded RDF data that is making a
statement about the resources available from that page. I thought
the connection between making statements about ethics could be made
similarly.
That is conceivable. I just haven't considered using RDF yet and it seems
to me that RDF would be more complicated.
Apparently I'm missing the idea of what you're doing with Nutch and
how you are representing ethical policies.
I'm not doing anything with nutch yet, I just have an idea that might be
useful and
I'm looking for a context for a reference implementation, which nutch
could be.
Bernhard
