Software Quality in Bioinformatics

The text is split in 4 parts:

Part 1 is an introduction to bioinformatics problems in general and it's the reason why I started writing in the first place.

Part 2 is a more refined list of software quality problems with specific examples from the bioinformatics community and a few generic principles to follow to increase quality.

Part 3 is the technical bit where I've tried to show a few practical examples on how to organize the development to achieve better quality.

Part 4 is the administrative final touch with some advices from a technical background to the management layer.

This text is not an attempt to create definite rules or state absolute truths of software development, nor I want to write another CMM, RUP, XP, Agile or such. These are just some bad practices I've found and what I've done (or would have done) to help increase software quality to ease maintenance, staff relocation and future developments.

Feel free to comment about anything (grammatical or technical) you want, I'd be glad to hear from you.

Update: I've recently attended a fantastic talk from Prof. Carole Goble (Univ. of Manchester) which express many of my concerns and much more: The Seven Deadly Sins of Bioinformatics.

1. Bioinformatics and Its Problems - For all audiences
2. Basic Principles - Reasonably technical, management friendly
   1. Standards
   2. Data on Demand
   3. Black Boxes
   4. Refactor or Rewrite
   5. Test Cases
   6. Document to Understand
3. Specific Problems and Solutions - Highly technical
   1. Code Organization
   2. Programming Languages
   3. Standard Libraries
   4. Patterns and anti-patterns
   5. Testing
   6. Optimization
      1. Architecture
      2. I/O Optimizations
      3. Profiling
      4. Parallelization
   7. Keeping Track
   8. Disaster Recovery
4. Policies and Standards Compliance - Management and administration friendly
   1. Quality Assurance
   2. Technical consortia
   3. Deliverables
5. Conclusions - For all audiences
6. Acknowledgements