# What you can do

> A shared grid with the formula library people already know, and columns that can hold a person, a task or a row of a dataset instead of text that looks like one

Source: https://aipril.dev/docs/spreadsheets/what-you-can-do

## The formulas you already know

The formula library is the one people learned in Excel - the same function names, the same
arguments, the same behaviour. Sums, lookups, conditionals, dates, text handling, statistics: what
somebody types out of habit works, and what they find in an old file can be pasted in.

This matters more than it sounds. The reason spreadsheets survive every attempt to replace them is
not the grid - it is that a large part of the workforce is already fluent in them. Keeping the
formula language means keeping that fluency, instead of asking people to learn a new expression
syntax to do what they could already do.

## Columns that mean something

In a file, every cell is text or a number, and anything else is a convention people agree to
maintain. Here a column has a type, and the type is enforced by what the cell actually is.

- **The ordinary ones** - text and long text, number, currency, percent, date and date-time,
  checkbox, rating, email, phone, link, attachment, barcode, icon, a code block
- **Choices** - single and multiple select, so a status column is a set of values, not whatever
  people typed that day
- **Computed and automatic** - a formula column, an autonumber, and the audit columns that fill
  themselves: created and last modified, and by whom
- **References to the rest of the platform** - an employee, an org unit, a position, a legal entity,
  a value from a dictionary, a row from a dataset, a task in a process, a linked record in another
  sheet, and a lookup that follows one of those links to show a field from the other side

That last group is the difference between a spreadsheet and a file that looks like one. A column of
employees is not a column of names somebody typed - it is a column of people, which is why it does
not go stale when somebody changes department, and why a process can route to whoever is in it.

## Working together

- Several people edit the same sheet at the same time, and see each other doing it. There is no file
  to send, no version to merge, and no "which one is current".
- A workbook holds several sheets, as a workbook always has.
- Comments attach to a cell or a range: a question about one number lives on that number instead of
  in a message thread somewhere else, with `@`-mentions that reach the person in their own channel.
- Rows carry actions, so a sheet can be worked through and not only filled in.

## Getting data in and out

- **Start from data you already have.** A sheet can be built from a dataset or a model: the rows are
  fetched and materialised into the sheet, and from then on it is an ordinary sheet you can work in.
- **Or start empty**, the way a plan usually does, and let what people enter become data.
- **What is entered is immediately readable** by the rest of the platform - see
  [Connected to the rest](https://aipril.dev/docs/spreadsheets/connected.md).

## Next

[Beyond Excel and Google Sheets](https://aipril.dev/docs/spreadsheets/beyond-excel.md) - where the line runs against the
tools people use today, and the two sheets you cannot build in Google Sheets.