Mac-OSX-Universal-Access

Contents

• How To Set The Cursor Size And Use Magnification Effectively
• • Cursor Size
  • Screen Zoom
  • Zooming Controls

How To Set The Cursor Size And Use Magnification Effectively

I did some reading about assistive technologies on Mac OS X for a friend of mine who has trouble reading screen fonts. The following are instructions on how to set the cursor size, and use magnification effectively. Please note that I am using Snow Leopard (10.6). If you are using Lion (10.7), or another version of OS X, you might notice some differences in the figures.

The relevant preferences can be accessed from the Apple menu found in the top-left corner of your screen.

1. Select “System Preferences” from this menu
2. Once it has opened, find and click on the “Universal Access” icon

   

You may also find a small “Universal Access” icon on the menu bar at the top-right of your screen. All you need to do is select the menu item called “Universal Access Preferences ..”.

There are a number of categories of Universal Access preferences: “Seeing”, “Hearing”, “Keyboard”, etc. First, I will show you how to adjust the cursor size.

Cursor Size

1. Click on the “Keyboard & Mouse” tab

   

2. Adjust the cursor-size using the slider bar

   

Screen Zoom

1. Next, switch to the “Seeing” tab

   

2. Make sure that “Zoom” access is turned “On”

   

3. In the “Zoom” section, locate and click the “Options…” button

   

4. The screen will look similar to this. For assistance when zooming, I recommend selecting the following settings:
   1. Set: “Maximum Zoom” to 3, and the “Minimum Zoom” to 0

      

   2. Select: “Zoom follows keyboard focus”

      

   3. Select: “So the pointer is at or near the center of the image”

      

Zooming Controls

When the screen is magnified and using the settings provided above, you will be able to effortlessly control the focal point of the magnifier with your trackpad, quickly moving from left-to-right or top-to-bottom of the full screen.

To get a bigger picture of what’s happening on the screen, I recommend you try zooming-out, and back in again. This is important to check for pop-up messages, and dialogs, or to locate windows that are outside the magnified view. The following two keyboard combinations provide a quick way to do this.

Zoom In	Zoom Out
option-command-=	option-command--

I encourage you to learn and use these keyboard combinations. Write them down on a sticky note and post it on the side of your screen or keyboard.

Visualizing-Open-Data

Contents

• A Recent Discovery
• Eight Years Ago

Related

• Sparse Matrix Collection

A Recent Discovery

As I was investigating software to visualize large graphs, I was delighted to discover the following image:

This was produced by sfdp, an algorithm from the Graphviz package that Yifan Hu designed to handle large graphs. Yifan has generated an impressive gallery of images from the entire Sparse Matrix Collection from the University of Florida.

One of the reasons I was happy to find this visualization is that I helped make it (eight years ago). The nodes represent instances of handwritten digits 0 through 9, and the edges connect similar instances. Shameless plug aside, credit is due to Yifan Hu for creating the software and also to Tim Davis from the University of Florida who created and maintained the public Sparse Matrix Collection, enabling scientists to benefit from shared data. I just fiddled another dataset (the canonical MNIST digits) and uploaded it as described below.

Another reason why I was glad to see this visualization is that it confirms the opinion that, when done well and used judiciously, visualizations of large graphs can be meaningful. In this case, visualizing the graph is meaningful because the separation of ten clusters in 2-dimensions is a proxy for their separation in a k-dimensional embedding. Since embeddings are commonly used to analyze relationships of similarity, we gain some insight into their effectiveness for a particular domain.

Eight Years Ago

I was working at a research institute in Boston for the summer. For a project there, I needed to benchmark our newly-developed, memory-efficient classification algorithm against the current state-of-the-art.

This boiled down to solving a large eigenvalue problem using the method as described in a corresponding paper. The problem was so large that we ran out of memory. This was back in the days of 32-bit addressing and we had maxed-out our memory.

Seeking advice, I wrote Tim Davis who is the author of the eigenvalue solver that was prescribed. He suggested that I submit this dataset to his Sparse Matrix Collection, and said he would get back to me. He did get back to me, rather promptly I might add, to say that his off-the-shelf solver would not be able to handle this problem, period. At the time, this was good news because it validated our decision to focus on memory-efficiency.