Detecting and Visualizing Inter-worksheet Smells in Spreadsheets
2 likes1,203 views
Felienne Hermans explores the common issues, or 'smells', found in spreadsheets by analyzing user experiences through interviews and applying software engineering principles. The study identifies specific spreadsheet smells, such as 'inappropriate intimacy' and 'feature envy', and reports the development of a visualization tool, Breviz, to aid users in understanding these complexities. Findings illustrate that concepts from software engineering can effectively inform the improvement of spreadsheet usage and design.
Detecting and Visualizing Inter-worksheet Smells in Spreadsheets
- 1. Detecting and Visualizing Inter-worksheet Smells in Spreadsheets Felienne Hermans (@felienne)
- 2. Code smells
- 3. How did we get this idea?
- 4. By studying what business does in practice
- 5. Excel is running companies!
- 6. 50% of spreadsheets form the basis for decisions
- 7. Spreadsheet users lack great tool support
- 10. We did not start coding!
- 11. We performed 27 interviews
- 12. We asked them: What annoys you?
- 13. And what makes you happy?
- 14. We found that especially support for understanding spreadsheets was missing
- 15. Then we started coding
- 16. Published at ICSE 2011
- 20. Users started diagnosing with the dfd’s
- 21. This led us to the idea of smells
- 22. We took Fowler’s smells and adapted them to be applicable on spreadsheets
- 25. Inappropriate Intimacy = number of couples between A and B
- 26. Feature Envy
- 27. Feature Envy
- 28. Feature Envy (calculated on a cell) =number of external connections
- 29. Middle man
- 30. Middle man
- 31. Middle man
- 32. Middle man = number of MMFs on a worksheet
- 33. Shotgun surgery
- 34. Shotgun surgery
- 35. Shotgun surgery =number of changing sheets and formulas
- 36. Then we set thresholds for the metrics Using the EUSES corpus as a bench mark
- 39. We validated our implementation with a case study
- 42. “Are those worksheets really that connected?”
- 43. “Are you sure that arrow is correct?”
- 44. “this formula annoys me, I have to go back to the other sheet so many times to look up the references, it makes me dizzy”
- 45. Interesting fact about Middle Man: Our definition was ‘wrong’
- 46. We had not anticipated Middle Men within worksheets
- 47. The most extreme case was a worksheet on which 220 formulas depended, spread over 10 worksheets
- 49. The owner immediately recognized the risk
- 50. Conclusion: principles from software engineering transfer well to spreadsheets
- 58. Detecting and Visualizing Inter-worksheet Smells in Spreadsheets Felienne Hermans (@felienne)