- Their clients requested their web page producing questionnaire software being more flexible, allowing multiple pages instead of one. A junior programmer finishes the task, the software get tested and rolled out. But - what a disaster - the saving didn’t function properly! Apparently there were no specs written, no documentation at all!
Since they do most of their development with no or incomplete specs my friend was looking for an easy way to jump start. Many people find it useful to have a template handy that helps them to categorize and jump start when faced with a problem they have to solve.
- Identify the problem
- Create a solution strategy
- Design a step-by-step plan.
- Proof of concept
- Is there room for improvement?
Is the problem understood and what is the real problem?
* If not then collect further information.
Is the problem definition clear and exact?
* If not then create a new and more precise definition.
Is there already a name for that problem?
* If not then create a short yet descriptive name.
Collect information about given input and desired output.
* Name and describe them.
Is this or a similar problem already known?
* Can you use or adopt an already existing solution?
Is there a solution for a less specialized problem?
* Is your problem a special case?
Can you split the problem in more simple problems?
* Start a new strategy for every partial problem.
Use your plan to solve the problem
Show that your solution solves
* in theory
* via empirical tests
Having some documentation could have helped avoiding the disaster.
Tags: apparently categorize descriptive name disaster empirical tests few days jump start jumpstart junior programmer new strategy partial problem precise definition problem proof proof of concept questionnaire software solution strategy