Aug 15
If some experiment is rerun, or you need to remove a column, start again from step 5. PS: My advisor drop by. I shown him the result and he didn't like it.
I think I'll buy some day these two books:
Vivid tables of experimental results in Latex
I've discussed with my advisor about how to show results in the tables for our eternal journal paper. We are comparing the performance of different algorithm/software. Should we bold the best times of each one? best time over similar problems?
I said the data is in the tables. So, bold should be use only for giving a suggestion of a fast impression. We finally agreed on using bold face for the harder problem of each family solved by each algorithm. By the way, on tables for Latex. As every latexer knows, it is painful to edit them on a pure text like emacs. I have tried everything, including Lyx and Texmacs, and then exporting them. Both are almost fine. However, I opted for total control by using Openoffice - Calc.- Collect your data. Summarize using lookup on sheets generated by your scripts, in case you want to rerun experiments. I generate many CSV and then import them into a document in Openoffice.
- Create sheets for each table to be used in Latex, linking what you want. Don't paste the data itself. Use links (special paste), so everything is keep update.
- Put columns with separators of line '\\', lines '\hline' and separator of columns '&'.
- Put multicolumn where you need, overwriting some '&' to keep the number of columns in sync.
- Finally, create a new document, copy the content of that sheet.
- Save it as CSV (Comma Separated Value) and using no separator or field. Every text on each cell will be concatenated with neighbor cells, but it's ok. You've used '&'.
- Keep the "save content of cell as it seen". So you keep rounding of numbers, etc.
Voilá!
You have wonderful Latex code you can paste in your tables.If some experiment is rerun, or you need to remove a column, start again from step 5. PS: My advisor drop by. I shown him the result and he didn't like it.
I think I'll buy some day these two books:
- The Visual Display of Quantitative Information, 2nd edition by Edward R. Tufte
- The Back of the Napkin: Solving Problems and Selling Ideas with Pictures by Dan Roam
