This question has been asked a few times but none that relate to a particular scenario: Graphical editors.

I've noticed several different patterns, mainly

  1. Inline Edititng - Complex nodes have several form-like elements which are edited directly on the object inline editing in Blender
  2. Nodes are simple, labeled objects wich can be selected. In selection, a (complex) property sheet or inspector tool window on the side is shown,which might contain many properties is shown and users can (like the Eclipse IDE does) enter image description here enter image description here

  3. A hybrid approach where nodes are quite simple (icons, shapes with labels) some properties (like the name) can be edited BUT on double click a pop-over or dialogs opens where users can edit all properties and then press commit/apply to finally apply the changes to the model. enter image description here

What I am looking for is the best approach given that my object are going to be of medium complexity (2 to 6 propteries), visually simple but very heterogenous from different domain models and I want to ensure a smooth user experience?

1 Answer 1


The purpose of such a workbench is to indicate relationships between various tiles and if required provide an option to position/edit tiles with reference to it.

Considering this primary purpose, I'd keep minimum information on the tile itself. The information on the tile should be limited to properties which are essential and enough to identify what the object is. Everything else can move to secondary properties window.

If you add too many properties on the tile, it will reduce the readability when there are too many nodes. Even if you add zoom functions, it will not be optimum.

I would ask myself, for the minimum set properties needed to identify each tile. I would also make use of shape/color variations so I have to keep less data in plain text. (Beware of color deficiencies and refrain from using only variations in color as an option)

  • So you'd advocate the 2nd or 3rd approach. Which one do you prefer? I would go for the hybrid approach (3rd), because essential properties are quickly accessible, but the user also can stay in focus.
    – Jack Shade
    May 28, 2015 at 12:59
  • Also you mentioned zooming functions. Consider this: When using semantic zoom, that is from far away nodes are simple, if you zoom in properties/tables/subgraphs appear making the choice above irrelevant.
    – Jack Shade
    May 28, 2015 at 13:00
  • I would go with the hybrid approach. I might have a small box at one corner, when I select a node, I can dynamically update that box with relevant properties. That's just me talking without knowing your precise use case though.
    – Harshal
    May 28, 2015 at 13:07
  • Also, I'm not talking about semantic zoom. I was just considering a zoom function if the whole workbench is complex.
    – Harshal
    May 28, 2015 at 13:09

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.