Zoom: Why two ways to do the same thing?

[pllim]

There can be only one.

astrowidgets/astrowidgets/core.py

Lines 409 to 427 in c594308

	@zoom_level.setter
	def zoom_level(self, val):
	if val == 'fit':
	self._viewer.zoom_fit()
	else:
	self._viewer.scale_to(val, val)
	def zoom(self, val):
	"""
	Zoom in or out by the given factor.
	Parameters
	----------
	val : int
	The zoom level to zoom the image.
	See `zoom_level`.
	"""
	self.zoom_level = self.zoom_level * val

[mwcraig]

I think these are two different operations, though arguably the names could be better.

zoom_level: a specific zoom relative to the scale (1 screen pixel) = (1 image pixel)

zoom: a factor by which to change the current zoom_level, e.g. "I want to zoom in another factor 1.5" . A better name might be zoom_by.

[pllim]

But do we need both? 🤔

[mwcraig]

I think we do in the context of user interactions. Imagine you have been doing some zoom in/zoom out by mouse or scroll wheel or whatever. You probably have no idea what your current zoom_level is but might know that you want to zoom by additional factor in or out.

One could calculate the new zoom_level manually but that seems cumbersome.

The other difference I see between them is that zoom_level is part of the state of the view/a property and zoom is an action/method.

[pllim]

So, should we remove the zoom_level setter and make it read-only?

[mwcraig]

I'm inclined not to because I see setting a specific level and zooming by a particular factor as different things. In a widget implementation, allowing zoom_level to be settable in the API means it is fine to make it a traitlet, which makes it easy to link the zoom level of two browsers.

[pllim]

Ah, gotcha...