Even better, follow the full process
1) Problem Identification
2) Requirements Capture
2a) Technology selection (optional)
3) Specification
4) verification and validation
Maybe you do some TDD and mix it up on 3 and 4, but this is the minimum process to meet an intent for a v-model, call it a shallow v.
Try and maintain at least minimum traceability thru the stages, but can easily fit each on a tab on a spreadsheet and link cells across the phases.
Once you get a few done, they are really very little work compared to the benefits yielded.
If you can't articulate what you need, want, how to do and how to check, even if only one stage in advance as you go, then you probably aren't ready to start.
Sometimes it might be instructive to write one to throw away if you aren't sure how it might best pan out, then make the plan for the actual thing.
No comments yet.
