recording each window change instead of a time interval

Create issue
Issue #53 new
Former user created an issue

First of all, thank you for your work, the idea of the program is great. However, I have a problem. From what I understand, the program takes a snapshot every 60 seconds. In a typical programmer's work, within 60 seconds we switch between many programs (code editor, browser, database manager). Very often switching is very quick to check coding results in browser. In such a case, the work on specific programs is not included in the working time. Shouldn't it be the case that every change of focus of the window is registered instead of these 60 seconds? (only the active window, others are not really needed for anything). Or maybe I don't understand something?

Comments (3)

  1. nomeata repo owner

    The idea is that in the long run, it evens out. If you peek at the browser for 10s at a time, then every 6th time will be accounted for, and counted as 60 seconds.

    I guess watching for focus change events would increase the precision. Unfortunately, I don’t think I will find the time to apply bigger changes to arbtt any time soon.

  2. Qcol

    in the long run, it evens out

    I don't think so. When I switch quickly between code editor and Chrome browser, in every 60 second, the work can look like this (let's assume such a case):

    • 50 seconds coding in SublimeText
    • 10 seconds checking the result in Chrome
    • 40 seconds SublimeText coding
    • 20 seconds checking the result in Chrome
    • 55 seconds SublimeText coding
    • 5 seconds checking the result in Chrome

    Every 60 seconds I can always hit Chrome which will show me that I worked 3 minutes only in Chrome without noting anything about SublimeText where I spent more time. It's just an example, of course, but it shows that we'll have wrong results throughout the day. Instead of targeting blindly every 60 seconds, it would be enough to register changes only when changing the focus of the window. I understand that you don't have the time, but this option is worth the introduction because the project itself with this small flaw is great. Greetings!

  3. nomeata repo owner

    With “in the long run” I mean statistically. This assume that there is no statistical correlation between your work and the second hand of the watch. Without such a correlation, if you spend 10% of your time using chrome, then ~10% of the samples taken will be while you use chrome.

    If you have a correlation, such in your example, it does not work.

    But I agree that the feature is worth considering.

  4. Log in to comment