# 7. Dragon ADR Position Set¶

• ADR - Architecture Design Records

## 7.1. Problem¶

• Set new position to x=10, y=20

## 7.2. Option 1¶

>>> dragon.fly(10, 20)
>>> dragon.teleport(10, 20)
>>> dragon.set_position(10, 20)
>>> dragon.set_position_xy(10, 20)

• Good: easy to use

• Good: encapsulation

• Good: easy to add validation if needed

• Good: easy to extend to 3D - add parameter with default value 0

• Bad: name set_position_xy() ties you to 2D point

• Bad: arguments are implicit, require knowledge of an API

• Bad: teleport() and fly() are bad names, does not indicate context what are the values provided as arguments

## 7.3. Option 2¶

>>> dragon.fly(x=10, y=20)
>>> dragon.teleport(x=10, y=20)
>>> dragon.set_position(x=10, y=20)

• Good: easy to use

• Good: arguments are explicit

• Good: encapsulation

• Good: easy to add validation if needed

• Good: easy to extend to 3D - add parameter with default value 0

• Bad: teleport() and fly() are bad names, to use-case specific

## 7.4. Option 3¶

>>> dragon.set(position_x=10, position_y=20)

• Good: easy to use

• Good: arguments are explicit

• Good: encapsulation

• Good: easy to add validation if needed

• Bad: set() is to generic

## 7.5. Option 4¶

>>> dragon.x = 10
>>> dragon.y = 20

>>> dragon.x, dragon.y = 10, 20

• Good: easy to use

• Good: arguments are explicit

• Good: can use @property for validation if needed

• Bad: names x and y are weakly related to dragon

Example:

>>> knn = KNearestNeighbors(k=3)
>>> knn.w = [1, 2, 3]


## 7.6. Option 5¶

>>> dragon.position_x = 10
>>> dragon.position_y = 20

>>> dragon.position_x, dragon.position_y = 10, 20

• Good: easy to use

• Good: arguments are explicit

• Good: can use @property for validation if needed

Example:

>>> knn = KNearestNeighbors(k=3)
>>> knn.weights = [1, 2, 3]


## 7.7. Option 6¶

>>> dragon.position = (10, 20)
>>> dragon.position @ (10, 20)

• Good: easy to use

• Good: can use @property for validation if needed

• Good: using @ (matmul) it is easy to validation

• Bad: @ (at) makes sense only in English

• Bad: require knowledge of an API

• Bad: not extensible, hard to refactor to 3D

## 7.8. Option 7¶

>>> dragon.position = Point(x=10, y=20)
>>> dragon.position @ Point(x=10, y=20)

• Good: easy to use

• Good: can use @property for validation if needed

• Good: arguments are explicit

• Bad: @ (at) makes sense only in English

• Bad: require knowledge of an API

• Bad: extensible, easy to refactor to 3D

## 7.9. Option 8¶

>>> dragon.position.x = 10
>>> dragon.position.y = 20

>>> dragon.position.x, dragon.position.y = 10, 20

• Good: more or less easy to use

• Good: arguments are explicit

• Good: can use @property for validation if needed

• Good: namespace

• Good: more or less readable

• Good: extensible, easy to refactor to 3D

• Bad: require knowledge of an API

## 7.10. Decision¶

>>> dragon.set_position(x=10, y=20)

• Good: easy to use

• Good: arguments are explicit

• Good: encapsulation

• Good: easy to add validation if needed

• Good: extensible, easy to refactor to 3D