Assumptions/Project Values

• Model-based/driven approach (via EMF)

  • Extensible

  • Vendor-neutral

  • Consistent view

• Initial support primarily (but not exclusively) for relational databases, yet design for non-relational

DTP Users (relative priority)

• Other Eclipse component developers (1)

• Developers

• Java application developer (2)

• Enterprise application developer

• Application architect (modeling, use-case construction, etc)

• DBA

• Report writer/business domain experts (3)

• Application architect

• Data architect

High Level Vision

• Connection management – generic

• Contributing companies:

• Sybase

• IBM

• Actuate

• ObjectNation

• Data sources/Data
  (includes mix of data connections and data forms)

• JDBC (Java developers)

• OLAP data sources (Report writer connectivity)

• XML

• File based: CSV, text files

• Applications (e.g. SAP, PeopleSoft)

• EDI

• Web Services

• Message Queues

• JCA

• Transaction management within connection management

• Read/write of application data sources

• Data source meta-models

• On-demand loading

• Model structure-based interaction online and offline

• Generate deltas

• Synchronization

• Object browsing

• Live view of the data

• Runtime Support (like SDO, EMF) (ability to run applications outside the IDE)

• In RCP

• Outside Eclipse entirely (non-UI only)

• Model type extensibility support

• Multidimensional

• Hierarchical (e.g. XML)

• Object-model (e.g. get data as objects aot tables and rows)

• Model transformation extensibility support (e.g. adapters to transform between one model to another)

General Use Cases/Requirements

• Design query against generic connection, read, generate generic report <BIRT>

• User: Java application developer incorporating reports into application; Report developer (e.g. generate sales reports)

• Create, edit, debug, deploy server component (e.g. Stored procedures, functions, …) <relational database>

• Acquire, modify, write, validate metadata from data source

• Query Support: Create, edit, visualize, debug, execute sql query

• Working with live view

• Change database table

• Add index

• Inspect data

• Lightweight administration

• User: Non-operational (for developers as opposed to DBA)

• Extensible

• Limited to well-defined interfaces (e.g. JDBC)

• Obtain/discover capability meta-data from a data source

• User: BIRT

SQL Development Tools Use Cases/Requirements

• Use connectivity project components

• Associate query with connection

• Create, edit, debug/execute, deploy SQL Statements

• Use common SQL Query Model/Parser

• User: SQL Query Tooling

• SQL execution plan support

• Retrieval, visualization

• Extensibility for optimization

• Support runtime parameterization

• Complex parameters

• Team support

• Q: how to persist SQL statements w/ associated connection

• Extensible persistence mechanism

• Multiple SQL statement support

• Create, edit, debug, save, deploy, etc of procedural objects

• SQL History view

• Result set support

• Model : SDO

• UI: extensible, handle vendor-specific data types

• Edit in place, reapply to database

• Handle large data sets

• Persist data results

• Multiple result set support

• Visual Query Builder

• Modular/componentized (e.g. separate components for Expression Builder, Conditions Builder, Joins Builder)

• Consider novice and experienced user

• Debug Support

• Vendor-neutral framework

• Multi-language support (SQL, Java)

• Extensible for debug support

• Extensible for UI presentation

Connectivity Break-out Session

There was some discussion about whether BIRT ODA should be in DTP or stay in BIRT. Later it was decided to move ODA into DTP.

Architecture

Layers

Connection Definition Layer (pure connections):

• Agreed to adopt Sybase connection management framework and port the WTP JDBC connection support to the framework.

• There was discussion about how ODA connections relate to connection profiles.

• ODA and the SQL model will wrap JDBC connections.

• ODA flat file, for example, would implement connection profile, driver.

Model Layer

• Agreed to adopt WTP SQLModel , SQLQuery, and DBDefinition.

View Layer

• Agreed to adopt the Eclipse navigator framework currently hosted in WTP project.

• Also need to work to have this navigator framework pushed into core Eclipse so DTP is not dependent on WTP.

• Agreed to use the Sybase contributed Enterprise Explorer as the instance of the navigator in DTP. The content provider currently hosted in the Database Explorer in WTP will be ported to the Enterprise Explorer for displaying database schema information.

Miscellaneous Use Cases

• Should support Import /Export Connection Profiles with password option (support BIRT report execution.)

• Connection sharing and transactions discussion.

• Read-only connections can be shared.

• Writable access would require “cloned” connections for transaction support.

SQL Development Tooling Break-out Session

Contribution Overview

• IBM Overview: SQL Model, SQL Query Model, SQL DML Parser, Data Definition Model

• Sybase Demo: SQL Editor and Debugger Framework

• Actuate: New Data Set Wizard

Component Identification

Infrastructure Components:

• Routines Editor Framework

• Routines Debugger Framework

• SQL Model/DB Definition Model

• SQL Query Model

• SQL Query Parser

• Common Connection Framework

• SQL Execution Plan Framework

Tooling Components:

• SQL Editor

• Base editor w/o persistence

• Content assist tied to model

• Multiple statement support

• SQL Query Builder (future)

• Results View

• Script History

• Database Explorer

Q: Should we think beyond relational data? If so, do you need a new higher layer such as ODA?

DTP Milestone Estimates

Subject to change based on further investigation.