To-Dos in Everyday Use
The scenario everyday of the to-dos has users who note down what they mean to do, and work it off. Its rules reckon their days in the time zone Europe/Berlin. Its figures are set by hand and coarsely, from what an evaluation of real to-dos showed, of which no figure is kept.
The Users#
- Users join one after the other over the range, and stay. On the day they join, they note their first to-do down and go about the rest of the day as about any other, so that even a data set of a single day shows more than to-dos noted down.
- Each user has an activity of their own: from about one to-do in two weeks for the least active to ten times as many, about one in a day and a half, for the most active.
- When users join and how active they are is spread evenly over the users rather than drawn at random, so that any population is a mix like any other.
- A user is about from 6 a.m. to midnight, Berlin time, mostly in the morning and again late in the evening. They are active on most days, and less on weekends.
- Once a day, a user goes through their to-dos. At other moments, they note new ones down.
The To-Dos#
A to-do is an errand: a verb and a thing it fits, such as Pay the rent, from the scenario's own lists, which name no people, places, or brands. About one in four carries a note in details that fits its verb.
About half of the to-dos are due on the day they are noted down, a quarter on the next day, and the rest in days or weeks.
When noting a to-do down, the user makes up their mind what they will do with it:
- whether they will put it off, which about half of the to-dos are, once at least, and a few again and again,
- whether they will adjust it, which changes its note two times in three, and otherwise its thing,
- whether they will mark it as important once it is due within two days,
- whether they will move it closer or do it early,
- and whether they will throw it away instead of doing it, which about one in twenty is.
Now and then, a user brings back a to-do they have just done or thrown away, by mistake.
What This Means for the Data#
- The population of the scenario is counted in weeks of a user: over a range of ten weeks, a user counts ten, and over a range of less than two weeks, one. A population that does not come to whole users ends with a user who joins correspondingly later, so the number of events grows about in proportion to the population, and even a few events over a long range come from a user who joins near its end.
- There is no history: a data set begins with the first user who joins, which may be a while after the beginning of the range.
- Nobody but its owner acts on a to-do, so the owner always knows its latest revision, and no command is ever refused.
- A single week of a user can already come out at more than a few events, so asking for only a few events, such as one or two, over a week or more is now and then refused with
range-too-long.