Often I come across the following situation: I have some experimental evaluation of running times of a new algorithm. Printing the running times in a table is good. Adding a plotted diagram illustrates the results in a nice way. Unfortunately this brings redundancy in a publication and needs many space, because the runnings times are already listed in the table.

Which way is best practice?

  1. Only print a table (precise results, but no illustration)
  2. Only print a plotted diagram (none precise results, but illustration)
  3. Both (precise and illustration, but redundancy)
  • 1
    Are the times important in and of themselves ("wow, he got it in 2s") or scaling with some other parameter ("wow, it goes as ln N")? I would think the former would favor tables and the later favor plots. – zeldredge Sep 1 '15 at 2:38

I'd go for the diagram, supplemented by a short table if need be. But be prepared to have a "full version" (with complete tables as appendices or so) as a technical report or otherwise available for anyone who wants to replicate your work (or use your data for comparison, or whatever).

  • This is a perfect use for supplementary information in a journal publication. – jakebeal Sep 5 '15 at 12:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.