The principle of timeboxing is simple. You decide how long you are going to spend on something, and that is how long you spend. No longer. If necessary, you modify the size of output or quality criteria to ensure that the task length does not over-run. This approach is very valuable in any situation where no adequate completion criterion exists.
If, for example, you are writing an article, or a report, then the completion criteria may include quantitative things like 'at least 3000 words' or 'commentary on at least three gothic texts'Â but there is no obvious qualitative threshold - you have no way of knowing when your essay is good enough. In this situation, time-boxing is useful for providing structure where none exists. You could, for example, say "OK, I have my 3000 words, now I will spend 45 minutes reviewing and improving my script. No more".
What timeboxing does is fix one variable - duration - and accept that another variable - quality, or completeness, or satisfaction - may vary. This differs from a traditional approach, in which quality-completeness-satisfaction are predefined and non-negotiable, but in which the duration of the task may vary considerably.