EDIT: The survey is now closed. Thank you all for your participation!
Greetings community
another month, another Micro survey!
This month we would like your feedback on the presentation of data in tables. We want you to be able to use tables efficiently without being confused by the presentation, as they are everywhere in Checkmk. So take a look at the survey to help us shed some light on the presentation of data in tables.
We recently received two support tickets criticizing the use of both scientific notation (e.g. 3.92e-3) and comma notation (0.392) in graphs and tables. We ran this micro survey to find out what’s important to our users when looking at data in tables and graphs, so that we can draw conclusions about data presentation formats.
We found out:
Clarity and ease of understanding (rank 1.7/4) and comparability (rank 2.0/4) are the most important characteristics for data presentation formats.
69% of all respondents agree that mixed data presentation formats such as 12.4 and 3.45e-10 are acceptable in certain circumstances: Otherwise compromised readability (32%), focus on precision (38%), and large differences between values (37%).
Mixed data presentation formats make it harder to compare values in a table, as peaks are more easily missed.
We concluded that we needed to make the mixing of data presentation formats more intentional. To achieve this, we are going to do three things:
Use the prefix of SI units (kilo, mega…) whenever it makes sense to make data sets more comparable.
Increase the threshold for using scientific notation for big numbers by 2. Currently, 12300 is written as 1.23e+4, which has more characters than comma notation. By raising the threshold, we ensure that there is a real benefit to using scientific notation: It is shorter.
Values less than 0.001 and greater than -0.001 are truncated to 0.000.
In order to incorporate the changes, we need to go through the metrics that currently allow scientific notation and define whether the SI unit prefix makes sense or whether the scientific notation is better. This will take some time, and as we are finalizing the development phase of 2.3, I can’t promise when it will be available.
This should not happen automatically, or at least there should be a way to get the required accuracy. At the moment I am dealing with some small numbers (without SI unit) and in my case I need at least a precision of 5 decimal places. These small numbers are added up over time and this should be visible in the graphs.
Thanks for the feedback @thl-cmk! Can you tell us a little bit more about the exact use case? (If you feel not comfortable posting it here, feel free to shoot me a private message )
Click here to start the survey 
)