Monday 5 December 2011

2011-11-30

Today, I finished implementing the first version of my primary and secondary tasks.

There are still some points to modify, especially the logging part. The original way in my program to write to a file has two main disadvantages. Firstly, it is hard to separate different participants' data, which increases difficulties in data analysis. Then, open and close files too many times waste time and decreases the speed.

Another important point to think about is how to give feedback of pen pressure to the participants.

1 comment:

  1. Young designers often overlook the cost of file and console input/output in their programs. RAM is several orders of magnitude faster than hard drive access, so buffering is generally desirable, especially when you have data coming in at a high sampling rate. Most operating systems will perform some buffering for you anyway, but managing your own buffer is better when you have a good idea of the volume and rate at which you will need to save data.
    Similarly, writing to the console (esp. command prompt in Windows) is quite slow. Having an efficient way to turn it off after debug helps your code run much faster. This is exactly what the Java logging API provides you with (http://docs.oracle.com/javase/1.4.2/docs/guide/util/logging/overview.html). It also allows you to format all your outputs in a central place, e.g. times tamping them all.

    ReplyDelete