[MOBY-dev] Notice that the ever elusive agent will start curating in 2 weeks

[Edward Kawas]

Hi,

The RDF agent that we have heard so much about will begin 'crawling' in 2
weeks starting Monday March 6, 2006.

As a service provider, your only task will be to generate one or more RDF
documents describing your service(s) and to place that document somewhere
reachable from the outside world. If you no longer wish to host that
service, then removing its description from the document will effectively
deregister the service.

The place to obtain your service RDF is from the following url:

	http://mobycentral.icapture.ubc.ca/servlets/forms/getSignatureForm

This page will update your service signature URL location.

In 2 weeks, the agent will start removing services that do not have a
signature url specified for it. The agent is also configured to remove any
service that fails to uphold to the latest API, as specified on biomoby.org.
Specifically, the agent is looking for un-named inputs or outputs (i.e.
simples or collections without article names). However, if a service was
modified, then the agent will use the API to re-register it and if the
service was invalid for other reasons, the registry will reject that
particular service.

In 2 weeks, the agent will attempt to read from the signature url that you
provided. If the agent cannot read from the url on 3 (maybe more or less -
details will be revealed later) different consecutive occasions, then the
agent will remove all services described by that url. The agent will then,
depending on configuration, email the service provider and detail what
happened to cause their service deregistration. 

If the deregistration was in error, service providers will be able to
reregister their services very easily. Contained in the email message to the
provider is RDF-XML that describes the last known footprint of their
service. Using this RDF-XML, and a tool to be released soon, you can
re-register your service in any registry.

As a service provider, you are free to add any valid RDF statements to your
document. At the very least, the document must contain statements with
predicates found at
http://biomoby.open-bio.org/index.php/for-developers/moby_extensions/moby_me
tadata. All other statements are ignored by the agent.

If you end up modifying the RDF-XML, and you want to verify that the XML is
valid, you could always use the following tool to validate and visualize
your RDF document:
http://www.w3.org/RDF/Validator/. If you are new to RDF and you do make
changes to the document, it's highly recommended to check your document with
this tool. 

A common error is placing multiple <RDF> </RDF> tags within a single file.
This is not allowed and will result in parsing errors. This error is usually
achieved by downloading multiple RDF documents and then placing the contents
of each document into one document. Those of you that wish to consolidate
RDF documents should consolidate the contents between the <RDF> </RDF> tags.


All questions and concerns regarding this agent should be posted to the list
so that they can be discussed amongst us all.

Thanks,

Eddie