Suppose I have a resizable graphic - users can zoom in and out. However, zooming does not scale the whole image the same way: some parts get zoomed in in a traditional way, but some may resize only in one direction (or even stay the same size as before). Example:

--------------                            --------------------------
|____________|                            |________________________|
      .          => when zoomed in =>               / \
 ------------                             --------------------------
|____________|                            |________________________|

The dot from the left becomes a big circle when user zooms in. While the dot scaled in x and y directions, the bars above and below scaled only in x direction, but not in y. Do you know of any research that looks at similar zooming behavior and how such zoom is perceived by the user?

  • 1
    Not a link to research, hence a comment, but I've seen this kind of behavior referenced as "semantic zoom" (as vs "optical zoom"). Microsoft are using this term for some of the zooming in Windows 8, but they didn't invent the term. Hope this helps someone's GoogleFu.
    – Bevan
    Commented Mar 21, 2012 at 23:12
  • 2
    @Bevan ,I think the link you are looking for is msdn.microsoft.com/en-us/library/windows/apps/hh465319.aspx
    – Mervin
    Commented Mar 21, 2012 at 23:29
  • Good link - but focuses on how, where the OP is interested in why.
    – Bevan
    Commented Mar 22, 2012 at 6:44

2 Answers 2


I'm afraid I can't directly answer the research part of the question, but I can rant about my personal opinion, for what it's worth: One long-standing problem I have had with zoomable interfaces is that they expand things rather than zooming.

The human brain is really very good at keeping track of where stuff is. It already uses a zooming interface and has done for centuries: pick up a piece of paper and hold it at arm's length, and then move it closer to read it. Congratulations, a zoomable UI. See how things get bigger in sync as they get closer! See how small stuff becomes readable!

If you do something else, you break the basic familiarity of a linear zoom. Try zooming out of a pdf and back in. It all makes sense, because your brain is already wired to understand it.

Since we are doing software not hardware, we have the option of making things disappear below a given zoom level, quickzooming between set levels and so on. These are good things.

In my opinion!


What features to scale when zooming?

The question is a little unclear to me. It totally depends on what features you got. I am not sure what kind of answer you are looking for here, or what resizable graphics we can suppose you are working with.

Optically, it's the content that zooms, and the controls that stay the same size. This is the case for photo editing programs, browsers or even Windows Magnifier. The question to ask yourself here is:

  • Where could the detail level need to be variable, and where is it better to stay the same?

But if you are looking for guidelines to make a semantic zoom, like @MFrank2012 links to in his comment, then I guess every zoom level makes its own special case depending on the content, and is probably developed using the technique:

  • If we put N elements of this type into a container of size X times Y, then how would we like the design to look?

You have to ask yourself what needs there are to zoom in your graphics, and then you will find what features to zoom.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.