MDSheet – Model-Driven Spreadsheets
Download as ODP, PDF0 likes2,878 views
The document discusses a project called MDSheet which takes a model-driven approach to spreadsheets. It aims to address common problems with spreadsheets by enforcing models and structure. The project has been ongoing for 8 years, looking at spreadsheets from both programming language and software engineering perspectives. Examples are given of how the model-driven approach can improve spreadsheets by reducing errors, improving understanding of data, and making the process more efficient overall.
MDSheet – Model-Driven Spreadsheets
- 1. Jácome Cunha, João Paulo Fernandes, Jorge Mendes, Rui Pereira, João Saraiva HASLab/INESC TEC & Universidade do Minho (rel)ease – Universidade da Beira Interior http://ssaapp.di.uminho.pt MDSheet – Model-Driven Spreadsheets SEMS '14 Delft, The Netherlands, July 2014
- 2. 13 History! › Project SpreadSheets As A Programming Paradigm (SSaaPP) › Working on spreadsheets' research for about 8 years › Programming languages perspective › Software engineering perspective
- 3. What Can Go Wrong? 1
- 6. 4 Do Models Grow on Trees?? VL/HCC'10
- 7. 5 But I Want to Change My Spreadsheet! VL/HCC'11 ICSE'12
- 8. 6 ICMT'12
- 9. 7 Did You Say Quality? ICCSA'13
- 10. 8 But I Don't Understand My Data... VL/HCC'14
- 11. 9 But... Does It Actually Works? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 -864 -691 -518 -346 -173 00 173 346 Pagamentos TimeDifference(s) 1 2 4 6 7 8 9 11 13 15 16 17 21 0% 20% 40% 60% 80% 100% Subject ErrorRate(%) 1 2 4 6 7 8 9 11 13 15 16 17 21 24 25 Su b je ct 1 2 5 7 8 11 13 15 18 21 22 0% 20% 40% 60% 80% 100% Model-Driven Plain Subject ErroRate(%)
- 12. 10 1 2 4 6 7 8 9 11 13 15 16 17 21 00: 00 05: 00 10: 00 15: 00 Su b je ct Time(mm:ss) 1 2 3 5 7 8 11 13 15 18 21 22 00:00 05:00 10:00 15:00 Model-Driven Plain Subject Time(mm:ss)
- 13. 11 DEMO
- 14. 12 Summary 1 2 4 6 7 8 9 11 13 15 16 17 21 24 25 00: 00 05: 00 10: 00 15: 00 Model-Driven P l ain Su b je ct Time(mm:ss) 1 2 3 5 7 8 11 13 15 18 21 22 00:00 05:00 10:00 15:00 Model-Driven Plain Subject Time(mm:ss)
Editor's Notes
- #4: Explain the example. Mostly, formula issues. But also incorrect values. It is also necessary a lot of work to add a new year or category. After adding data, it may be necessary to update the formulas.
- #5: Our solution it to have a specification of the spreadsheet one wants and make the spreadsheet follow it. It helps the user to keep the data correct as only the input cells are allowed to be changed. The formula are automatically updated in whenever necessary. To add a new year or category just press one button
- #7: No they don't!! We infer them from data instances, using data mining techniques. A relational model is computed, and from that a ClassSheet is obtained.
- #8: Change the model to an evolved one. You can edit the model and have the data automatically co-evolved.
- #9: You can also evolve the data and have a new model for it. Why? Because it is probably more natural for some people. Also, because some operations are also easier to do on the data than on the model.
- #11: You can Query it!!
- #13: Begin query construction End query construction
- #15: Spreadsheet tend to have errors We have devised a model-driven approach to help users Models can automatically be inferred And guide the user in introducing correct data We have empirical evidence that our system indeed helps users to be more efficient and effective