1. Distributed development, layers of frameworks, and Java’s modeling
limitations make it easy to create bloated data designs.
2. In this environment, awareness of costs is essential for making informed
tradeoffs.
3. Some tradeoffs can be made without sacrificing speed or sound design.
4. The concept of data structure health – the ratio of actual data to its
representation – can illuminate where there is room for improvement,
and point out aspects of a design that will not scale
(Full Story: Building Memory-efficient Java Applications – IBM Researchers)


February 2, 2012
