[OpenSPP-project] Notes from Team meeting of 20110616

[Dean W. <dea...@so...>]

Attending: Vijay, Jean Francois, Jeremy, Dean.

Notes by Dean

Summary: Code for each developer group is coming together, but has not yet been integrated into a deployable whole. We are shooting to be able to demo the three basic transactions on Destination Group on the June 23 call. This means we'll need to start pushing integration over the weekend.



Status report: Vijay

Created data model for destination group and checked in code for it. Created Data Access Object for those transactions with database persistence. Next step is to get it running under the Tomcat container. Also need to verify compliance with coding guidelines. open question on transforming string identifier to numeric ID. Now doing a database lookup for this.

Need to work more with Jeremy on package names and getting into deployment server.

Discussed string-to-number mapping. Probably need to have a schema walkthrough technical meeting. For example, which identifiers must be unique? Which ones do we use as primary or composite keys? Some of this can push to next sprint.


Status report: Jeremy

Samuel and Mikkael have been working on CXF for service environment. They have generated out a service stub. They are having a  deployment issue with the namespace of the endpoint; not finding the endpoint under Tomcat. They are also using SoapUI  to test. Jeremy is working on the command processing logic. he will shortly send out some emails on package names and package contents. Also will start working on exception architecture, which has to encompass the Spring exception model.

Mikkael and Samuel are working on Log4J usage and need to nail down some conventions for usage.

Issue: Coding guidelines and comments in documentation. Some of the code is generated by a tool; this could does not conform to the guidelines.   Some of this code can be dynamically generated from Ant, other of it requires hand-modification post-generation and therefore cannot be dynamically generated and will be checked into SVN. Jean-francois asked Jeremy to send out an email explaining this so that CableLabs can approve this variation from contracted coding guidelines.

Discussion of packaging: Proposed that we could put everything in one big JAR for now. More discussion will follow.

Discussion of exceptions followed. Proposed that the business logic layer should catch all lower-layer exceptions and turn them into something reasonable for the client. This raises questions for logging of exceptions and will need to be dealt with as part of Log4J process.

Question: Does protocol spec provide for strength-length limitation responses?


Dean reported on the dev/test server. We need to define process checkpoints for moving onto the server.

Discussion of Process: Ant has targets that check out code from SVN, onto build server, run through deployment scripts, create databases, etc.  Noted that dropping the database between releases may require repopulating the database with reference data. Vijay proposes database core scripts in SVN so that we can rebuild quickly.

When do we expect to start putting "stuff" onto the dev/test server? Need to get service layer, command layer, DAO, schema, and Ant. Can we have demonstration by next Thursday? This means we need to start trial integration Sunday.

Vijay and Jeremy to talk at 0730 EDT Friday. Topics: conventions, logging, exceptions, how we see integration coming together, and design questions around commands and CXF, DTO, DAOs, ANT scripts.

Question: What does the demo look like? Envision using soapUI  running shared-screen from GotoMeeting or (something similar). Demonstrator will show the add, get, delete, and get(not found) operations on a single destination group.