Hello, we are evaluating the possibility to use Slide for huge amounts of documents.
One needed feature would be to support folders with millions of documents. It's clear that someone could not navigate into such a folder anymore but would be restricted to search (via DASL) before he could have access to a document. Our current plan is to implement our own store and our own search engine. The backend would be a RDBMS. But before we start we want to make sure that this can work. So we studied the source code a little bit. One thing we stumbled over is the following: Currently the class ObjectNode is used as a parameter in many places. For example "NodeStore.retrieveObject(Uri uri)". ObjectNode is an abstract class with protected member variables "children" and "links" both of type "Vector". Our plan was to implement a "lazy" ObjectNode subclass which would not store the list of children in memory but create it "lazy" during access. So the question is: what's the reason to force subclasses of ObjectNode to use Vectors ? An alternative would be to not use member variables inside ObjectNode but only interfaces. Another question is: why do you not use the newer interfaces of the collection framework (List,Iterator) instead ? The third question: is there a chance that a patch that would change this would be accepted ? Ciao, Olli -- Oliver Geisser Ceyoniq Technology -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
