ORCSolver: An Efficient Solver for Adaptive GUI Layout with OR-Constraints

Conditionally accepted by the ACM Conference on Human Factors in Computing Systems (CHI), 2020.

Authors: Yue Jiang, Wolfgang Stuerzlinger, Matthias Zwicker, Christof Lutteroth

Paper: http://www.cs.umd.edu/~yuejiang/papers/ORCSolver.pdf

Video: https://www.youtube.com/watch?v=0S77vVG8btE&feature=youtu.be

Related Paper: ORC Layout: Adaptive GUI Layout with OR-Constraints

Paper: http://www.cs.umd.edu/~yuejiang/CHI2019/paper.pdf

Video: https://www.youtube.com/watch?v=eiEmLTfPDZQ&feature=youtu.be

To get started:

git clone https://github.com/YueJiang-nj/ORCSolver-CHI2020.git

Prerequisite installation

1. Python3 
2. Tkinter (GUI package)
3. CVXPY (a Python-embedded language for convex optimization problems)
4. Microsoft Z3 Solver [Optional] (only needed for running Z3 version)

Layout

The project has the following file layout:

README.md
API_ORCSolver.pdf
Code/
  ORCSolver (Ours)/
  PureBranch&Bound/
  PureZ3/
  QPforFlows/
  images/

The API_ORCSolver.pdf includes the API of ORCSolver which can be used to specify layouts. Our API allows us to plug in different solvers for different ORC patterns (including ORCSolver(Ours), QP for Flows, and Pure Branch & Bound).

The Code folder contains all the source code for all the four methods as mentioned in the paper (ORCSolver(Ours), Pure Z3, QP for Flows, Pure Branch & Bound). We also provide sample code for some layout patterns and the code generating examples in our teaser and video.