Janklab Goals

These are the goals guiding Janklab’s development. In addition to goals, we explicitly define non-goals, the things we are willing to give up in order to achieve our goals. Engineering is all about trade-offs.

Goals

  • Concise client code
  • Portability of non-OS-specific features
  • Support for useful OS-specific features
  • Support for polymorphic programming
  • Avoiding edge cases and “discontinuities” in our data types

Secondary goals

  • Concise Janklab implementation code

Non-goals

  • Modularity
  • Clean top-level namespace
  • All features being fully portable across OSes
  • Back-compatibility of new Janklab releases with older Matlab releases
  • Minimizing third-party library dependencies
  • Small distribution file size
  • Minimizing number of classes and functions
  • Running without Java

MDBC

Goals for MDBC

  • Support major open database vendors
    • PostgreSQL, MySQL, Microsoft SQL Server, Oracle, Derby
    • Major criteria: a free-as-in-beer downloadable version of the DB server for development and testing
  • Placeholder parameters and prepared statements
  • Efficient large INSERTs
  • Date/times as real date/time types
    • Definitely not strings, and not raw numerics/datenums
  • Smooth edge cases and uniform types
    • E.g. empty result sets returned as empty table types, not ‘[]’
  • Support “INSERT … RETURNING”
  • Support JDBC driver tuning parameters
    • Preferably with reasonable Matlab-tuned per-flavor defaults
  • Symbol/categorical support with Java-side conversion
  • Bonus: extensibility for DB extensions
    • Like PostGIS and user-defined types
  • Bonus: NoSQL databases that present PostgreSQL or other JDBC APIs

Non-goals for MDBC

  • Harder-to-work-with or non-free databases
    • IBM DB2, Sybase
  • Non-JDBC/non-Java drivers
  • ODBC support
  • Drop-in compatibility with base Database Toolbox