Bottom-Up Design for Lab Report:
Conclusions first

Kris Stewart, CS 575 Supercomputing
Fall 2003

Courses such as RWS 503W Technical Writing (Technical Communication: Situations and Strategies, Mike Markel, St. Martin's Press, 1996) develop the skills over a semester to design, write and present professional documentation. This is an excellent course that I would recommend to all Computer Science students for satisfying their Upper Division Writing Requirement for GE and tends to follow the Top-Down Design approach for documentation preparation. Within CS 575 you have course-specific constraints: 1) documents must be readable on a UNIX system, 2) documents are the Lab Report reporting the results of a Computational Experiment from a program that you have written (modified). Therefore, the authors of the Lab Report also control the mechanism used to generate the data which the Lab Report documents.

Development Cycle for Computational Experiment and Lab Report

A Learning through Discovery approach that has been successfully used in this course is:
  1. Adapt the sample code to explore your problem
  2. Use Programming Exploration to gain familiarity with the data that can be generated for your problem.
  3. Refine your program to focus on conclusions for your problem.
  4. Identify conclusions and make them clear
  5. Write body of report to develop terminology used in the Conclusions.
  6. Write introduction of report provide a succinct statement of the problem you are investigating and what you will cover in your Lab Report. You want to entice someone to read your report.

I call this Bottom-Up Design because the initial focus is on discovering what conclusions you can make based on the data your program produces. Computational experiments are unique since the student is not given a step-by-step algorithm of what to accomplish. Instead you are given a scientific problem to examine and you are asked to use the knowledge you gained from earlier experiments to establish convincing arguments to support your Scientific Problem solution. The student controls the code and you should feel free to modify it to DISCOVER more. You can then modify your program to produce nicely formatted data, to support the conclusions that you have decided on. As you obtain additional data, you just might discover more information about your problem.

The Order for Writing the Lab Report

When you approach the actual writing of the Lab Report, it should follow our standard format of: a) Introductory Section, b) Narrative Section, c) Concluding Section, d) Appendix (with well-labeled data tables). But take care to keep your writing focused on the conclusions you have discovered. Therefore, I recommend that the first section to write is your Concluding Section.
  1. Appendix (with well-labeled data tables)
  2. Concluding Section
  3. Narrative Section
  4. Introductory Section
The Concluding Section summarizes the careful arguments you present in the Narrative Section. The Narrative Section should not ramble, instead it should be focused on stating the problem, developing the terminology that you use to make your conclusion and establish the specific linkages with your data (in well labeled tables) to support your conclusions. Since you will have conclusions on three separate items (the scientific problem, the work performance, and the accuracy performance), the body of your report should develop these ideas in separate, focused subsections. Your data should be referred to in the Narrative Section by Table Number so that the reader can clearly identify the data that convinced you of your conclusions. It is often helpful to provide a summarizing table in the Narrative Section, pointing out where in the Appendix the original data is found. The final section that you actually write is the "Introductory Section" to entice the reader to read your Lab Report.

The Order for the Reader of a Technical Paper

In the technical world, it is often the case that the first section to be read is the Concluding Section, to see if there are some worthwhile, interesting results. If the report looks interesting, then the Introductory Section is read to understand what the report is actually covering, then the Narrative Section, then the Concluding Section.

Once you have put a lot of effort into a project and worked hard to understand new concepts, it is tempting to include everything that you learned in your final report. But this can lead to a lengthy document that dilutes the focus on the really interesting results that you worked so hard to obtain.