Monthly ArchiveFebruary 2009
Databases & Habanero 20 Feb 2009 06:14 pm
Hell is testing without an in-memory database
Application maintenance tends to be a cumulative experience. Each new bug report, however inconsequential, comes attached with increasing levels of resentment from the development. You fix, you create patch, you launch live, you accidentally break something else, you get phone call. Customer unhappy, developer burnt out. Developer changes jobs, application is passed around like a hot potato, and the next developer is even more likely to break things than the first, given their lack of knowledge of the application’s “unique problem solving strategies”.
Continue Reading »
Habanero 17 Feb 2009 04:34 pm
Habanero v2.2 alpha is Live
A look at the new features list could make the v2.2 release look relatively minor. This is far from the truth. The relationship level has been almost completely rewritten, now supporting three variants of relationships: association, composition and aggregation.
The rewrite was born out of a continual conflict between expectation and implementation. Let’s look at an example to clarify where the confusion comes in. You have an Order and its OrderItems. It’s easy to understand that there’s no reason for an order item to exist if there is no order. Effectively, they’re one complete unit - you’d even imagine that an Order is useless without order items as well. The trouble with coding this kind of relationship is that your mind tells you you’re working with one conceptual unit, but you’re effectively coding for two.
Continue Reading »
