Kale: A Spreadsheet System That Prevents Table-Editing Bugs
Kale: A Transformation-Safe Spreadsheet System
Spreadsheets often break when users insert or delete rows or columns in referenced tables, because formulas automatically adjust to new ranges that may not match what the user intended. A user study by the authors confirms that these standard reference semantics are error-prone and pose significant risk. They introduce Kale, a prototype that eliminates such bugs by restricting the kinds of references allowed. Tests show users can complete tasks effectively with Kale, and a corpus study assesses the impact of these restrictions on real-world usage.
When a user changes the structure of a referenced table, the spreadsheet system updates the references to refer to a new range, but this new range may differ from the user's expectations, introducing bugs in spreadsheets.
- mrbluecoat
At least I won't be tempted to eat it.
- TacticalCoder
> Spreadsheet formulas can refer to rectangular ranges of arbitrary size. When a user changes the structure of a referenced table, the spreadsheet system updates the references to refer to a new range. Unfortunately, this new range may differ from the user's expectations, introducing bugs in spreadsheets. We describe a user study showing that standard reference semantics are error-prone
Yeah. It's a pet-peeve of mine: after having been tasked to port a spreadsheet to a dedicated app, I've lost all illusions about spreadsheets.
Or, put it like this: spreadsheets are for programmers. When a non-programmer uses a spreadsheet, it's full, full, full of broken assumptions.
The famous Excel vid by Joel Spolsky shows just that: it's ultra-powerful stuff, but you basically have to be a programmer to understand what's doable and how to use that power wisely (and Joel Spolsky is a programmer himself).
Users don't understand dates and the relation between values and the values at a certain date or the values at the date of encoding or the values when they're automatically fetched, for a start.
Then they have absolutely zero clue as how to model things crossing boundaries: for example something crossing a year boundary.
I could go on and on but it's just mind-boggling. Reunions discussing the spreadsheet with "the users" and asking questions and they're just lost: they don't get it. They don't understand what they've done. It's confusion everywhere (in their minds and in their spreadsheet […]
- krautburglar
Since free relational DBs like sqlite & mysql came along, only optimal use I see of spreadsheets "in the wild" now is psychological (i.e. real-time "massaging" of the numbers to overcome objections). For things that are more objective, we have so many superior options. Not just databases either. Jupyter & various other REPLs are also great.