On 10/10/11 12:25 PM, Richard Eckart de Castilho wrote:
When you say "per type system", I assume you mean "per project"? Imagine one
uses the proposed CAS Editor plugin which can resolve type systems from the Classpath of a Java
project, then the visible annotation types would be stored for that Java project, right?
No the scope is really per type system, and not per project, since a
project can contain multiple type systems.
The Cas Editor Ide Plugin creates a preference store file per type
system, and then saves all related settings to this type
system there. This is for example the styling settings of the
annotations, project teams usually want to share this file.
This will likely be extended with more settings, such as annotation type
to key mappings.
The OpenNLP Cas Editor plugin will store there configuration which is
related to the type system, e.g. which annotation types
should be used to extract Sentences and Tokens.
Do you suggest to only have a persistent per type system storage which
stores all settings, or do you also believe that some
settings should only be "remembered" per session.
Depending on the usage the per type system preferences might just be
stored on a server, that is for example I am planning
to do for the Cas Editor Corpus Server integration. Then a team will
automatically share this settings file. Or should that
better be local (not shared)?
Jörn
