Evaporation Laboratory

Manual

Kenneth D. Forbus

Qualitative Reasoning Group

Institute for the Learning Sciences

Northwestern University


Version of May 7, 1997

In case of bugs, please send email to

evaplab-bug@cs.northwestern.edu

1. Introduction 12. Installation and Troubleshooting 12.1 Troubleshooting 23. Starting the program 24. Running a simulation 55. Inspecting the results of a simulation 55.1 Info mode 65.2 Hypertext 65.3 Behavior Summary 75.4 Graphing data 76. Some suggestions for using the laboratory 87. Exiting the program 88. Acknowledgments 8

  1. Introduction

The purpose of this software is to enable you to experiment with the phenomena of evaporation. Evaporation is important in its own right - for instance, if you understand evaporation you can figure out why a healthy dog's nose should be cold and wet, and why the ocean's average temperature hasn't changed much during the history of the Earth. Understanding evaporation is also important for understanding how weather works, since evaporation is how water gets into the atmosphere. Watching evaporation happen, however, can be a bit dull - watching paint dry is literally an example of evaporation at work. With a good computer several simulated hours can pass in less than a minute, making it easier to experiment with evaporation.

Here is what the simulator is about. You have a set of cups, all alike in shape and size but made of different materials. You can put varying amounts of water in these cups, and vary the temperature of the water. You can place the water in one of a number of places, such as Chicago or Las Vegas, and come back in a few hours to see what happened. The software automatically monitors certain parameters, and provides tools for graphing these parameters and asking questions about them and about the simulation in general. You can create several simulations, trying different things out to see what happens. Results from your explorations can be saved in a disk file, for further contemplation or for inclusion in a report.

  1. Installation and Troubleshooting

The software comes in one of two ways:

  1. A set of floppy disks
  2. A large Zip archive

If you receive a Zip archive: Use your favorite extraction software to unpack it, making sure that you enable it to create subdirectories. Unpacking it yields several directories, called DISK1, DISK2, etc. which contain the contents of the installation floppy disks. Copy each of these directories to a floppy disk, labeling it with the name of the directory it came from (i.e., the floppy containing the files from the directory DISK1 should be labeled DISK1). Then follow the instructions below for installing from floppy disks.

If you receive floppy disks: Insert the first disk in an appropriate drive, and run the program SETUP.EXE on that disk. This starts an installation program that leads you through the rest of the process. If you don't know how to run a program from a floppy disk, please read on.

To run SETUP.EXE: Make sure DISK1 is in your floppy drive. If you are using Windows 95, you can either double click on the icon for that drive, then double click on the icon for SETUP.EXE, or click on the Start menu, click on RUN, and type A:SETUP.EXE. If you are using an earlier version of windows, you can either use the File Manager to look at the floppy and double-click on SETUP.EXE, or go to the Program Manager and click on File|Run and type A:SETUP.EXE.

  1. Troubleshooting

This software has been tested under Windows 3.1, Windows for Workgroups, Windows NT, and Windows 95. However, it is impossible to guarantee that it will work under every possible combination of hardware and software. You need a video system capable of at least SVGA resolution, with 256 colors. If you run at very high resolution (i.e., 1600x1200) you will have trouble installing the software, although it may run correctly once you have installed it. The software will not run in video modes deeper than 256 colors.

If you are using Windows 3.1 or Windows for Workgroups, and the software either does not run at all or halts immediately complaining about missing library entries, the most likely cause is that you either do not have the Microsoft Win32s extensions installed, or have an outdated version of them installed. This software requires that you have version 1.30 or later of the Microsoft Win32s extensions installed. This software can be found at Microsoft's web site for downloading or from your local systems administrator.

  1. Starting the program

The installation script creates a program group (if you are using Win 3.x or Win NT or Windows for Workgroups) or a start menu item (if you are using Win95). To start the program, select the icon/item as you would any other windows program.

The Evaporation Laboratory has several windows:

What these windows look like is illustrated in Figures 1 and 2. You use the Laboratory by setting up and running simulations, inspecting, comparing, and thinking about the results. You can create a number of simulations, and save the results to files. Running multiple simulations enables you to see the consequences of changing something. Comparing the results of two simulations can be a powerful way to figure out what is really happening.










Figure 1: The setup window



Figure 2: The inspector window
  1. Running a simulation

To run a simulation, you must be in the Simulation Setup window. You get to this window by clicking the Simulation Setup button. Here are the steps you need to do in order to carry out a simulation:

  1. Select a name for your simulation, and type it in the Simulation Name field of the form.
  2. Choose a cup from the catalog of cups. Clicking on the label Cups to put water in gives you a catalog of cups to choose from. You can browse this catalog by clicking on the down arrow to see the alternatives. As you change the selected cup, the picture and descriptions change accordingly.
  3. If you want to customize the parameters of your cup, use the tabs to access the variables that can be manipulated (in this case, the mass and temperature of the water in the cup). Be sure to click Set if you want the new value you have chosen to be used. If you don't click Set, your changed setting will be ignored.
  4. Click on the Environments to experiment in tab, and proceed as you did in steps 2 and 3 for selecting and customizing a cup. (One can select catalog entries in any order, of course, and can change one's mind about selections until the simulation is actually run).
  5. You can change the duration of the simulation by clicking on the hours/minutes/seconds clock in the upper right of the window.
  6. Once a selection from both catalogs has been made, start simulating by clicking the GO! button.
  7. The progress indicator shows you how far the simulation has proceeded. Four hours of simulated time on a 90 MHz Pentium takes less than 30 seconds. When it is finished, the program brings up the inspection window so you can analyze your results.
  1. Inspecting the results of a simulation

There are several different kinds of things you can do with the Inspector window. You can

You can do these things in any order you like. The inspection window is divided into three parts. There is always a set of buttons at the top that let you set up new simulations, save files, and exit. The rest of the window is divided into two smaller windows, each of which can be used to display information about simulations. There are three ways to use these subwindows, which you control by your choice of Mode. These modes are

Let us look at each in turn.



  1. Question/Answer mode

Question/Answer mode enables you to examine several kinds of information about a simulation. You can look at

Each of these actions changes the contents of the Things to ask about box, the Description box, and the picture displayed. As its name suggests, clicking on items in the Things to ask about box provides a way to ask more questions about a simulation, using the hypertext facilities of the Laboratory. This explanation facility is a bit unusual for simulators, so let's examine it more closely.

  1. Hypertext

Many of the parts of the Inspector are hypertext. What that means is that if you click on something, you may get a menu of questions that can be asked about that thing. For instance, if you click on a line that says

evaporation from cup

a menu appears that contains two questions. In this case, the menu has two questions, what can evaporation affect and when can evaporation occur. If you choose to ask what evaporation affects, the answer you get is

mass of water in cup

heat of water in cup

mass of vapor in Atmosphere

heat of Atmosphere

This does not tell you all the indirect effects of evaporation, of course - but clicking on any line of the answer provides yet more questions that you can ask to explore this issue. This explanation also does not tell you which way evaporation affects each parameter: does it cause it to go up or down? If there are two effects on something, which one wins? Questions like those you will have to answer by setting up simulation experiments and analyzing the results.

If you click on something that doesn't have any questions associated with it, either nothing will happen, or a small window saying No Questions will appear.

Sometimes the list of questions gets too long for comfortable reading. In that case, use the Clear Questions button to clear the explanation window's contents. It is often useful to save the questions and answers you have generated as a text file, so that you can think about it more and perhaps include it in your notes and/or reports. The Save button lets you select a file for saving this information.

Most lines in the explanation box are hypertext, in that they have questions associated with them. Clicking on them is a good way to ask follow-up questions. How does one start asking questions in the first place? Every other part of the inspector provides topics that can be starting points for questions: The behavior summary, the simulation initial state information, and the graphs. Let's look at those next.

  1. Behavior Summary


The behavior summary is a hypertext that describes, in English, how the system behaved during the simulation. It describes when important entities exist (i.e., the water in the cup) and when certain important relationships hold (i.e., that the water in the cup touches the atmosphere). It also describes what physical processes are acting in the situation. Time in the simulation is divided into intervals when these important properties change.

If the simulation consists of a single interval, then everything that existed before remains in existence and whatever processes were acting on the situation continue acting on it during the entire simulated time. This does not mean that the parameters themselves are not changing! For example, if we watch someone filling a swimming pool, our description of what is happening ("water is flowing into the pool") doesn't change even though we know that the level of the water is rising.

If the simulation consists of more than one interval, then some important change happened during the simulated time. If we looked at a simulation of a swimming pool being filled, we might see a description like

Between 0.0 and 2.1 hours the following occurred:

flow of water into the swimming pool

there is water in the swimming pool

Between 2.11 and 4.0 hours the following occurred:

there is water in the swimming pool

The difference between the two descriptions tells us what happened at 2.1 hours: For some reason, the flow of water into the swimming pool stopped. The lines prefaced by arrows are hypertext, so you can click on them to ask questions that might help you figure out why some change occurred.

  1. Graphing data

There are two modes for graphing. Graph mode lets you look at one parameter of one simulation at a time. You can get additional information by clicking on the graph. Clicking with the left button displays the numerical value of the graph at the time you clicked. Clicking with the right button displays a menu of questions about that parameter, just as if you had asked about that parameter through the simulation catalog.

Clicking left on the graph provides the exact data and time for the corresponding point in the graph, whereas clicking right provides a menu of questions that can be asked about that parameter. These questions use the hypertext system to display the answers.

Notice that you can look at two pieces of data at a time by putting one graph in the top window and the other graph in the bottom window. It is often very useful to compare the same parameter across two simulations, both to see how it changes over time and how two simulations end up being the same or different. Multi-Graph mode lets you do this. You can choose a parameter and two simulations, and everything is scaled to fit on one graph. You must have carried out at least two simulations before you can use multi-graph mode, of course.

  1. Saving your work

You can save two kinds of information about your simulations:

  1. Numerical data. You can save the numerical data into a file in a format called "comma delimited form", which is readable by spreadsheets and many other programs. This is useful for creating different types of graphs, printing reports, or analyzing the data further.
  2. Question/Answer sessions. You can save the contents of your question/answer sessions with the simulator into a text file. This file can then be printed or pasted into your reports.

To save data, press the Save button, and supply the information about which kind of information to save, what simulators to save it for, and where to save it that the dialog asks for.

  1. Some suggestions for using the laboratory

Using a simulation to come up with a theory explaining what is happening and why requires doing experiments. Sometimes experiments are exploratory, just trying things out to see what happens. Sometimes experiments are designed to test specific hypotheses, such as finding out exactly how one parameter affects another. You may find that comparing two simulations can be very helpful in figuring things out. Here are some ways to use the Laboratory in making such comparisons.

  1. Exiting the program

You can exit the program by clicking the Exit button in either dialog. The program will ask you if you are sure, especially if you have created a simulation whose results have not been saved.

  1. Acknowledgments

The development of this software was funded by DARPA under the Computer Aided Education and Training Initiative (CAETI). It uses ideas developed through basic research funded by the Office of Naval Research. The C++ runtime system was written by Aaron Thomason, John DeMastri, and Mike Oltmans. Penelope Sibun created the descriptions in the environment catalog. Katarina Obradovic generated the images of the cups and other associated artwork.