Hi all,
We are building a bespoke application which is storing documents in an Alfresco repository through a purpose built user interface. The presentation layer of the application is being coded in JSP and we have successfully written pages which insert and retrieve documents via calls to the Web Service API. Initially I actually had the java embedded in the JSP and have now abstracted it out to Java Beans but still runnning within the context of the servlet container. It works ok, but it is a bit sluggish and I think this is mostly down to the fact that it is essentially stateless code and is repeatedly logging into the repository and back out again.
Is there anything to stop me implementing a server of my own in either an EJB container or possibly CORBA using the Apfresci foundation API but also having the Alfresco web client running in Tomcat at the same time against the same repository? That way, I can make my bespoke server stateful and cut out a lot of the latency. Initially, I assumed that there would be locking issues and conflicts, but on further reflection it seems like the web client must work in pretty much this way since it is running in a servlet container rather than a full apps server - am I wrong?
many thanks in advance
Martin