Introduction

Welcome to CardSword: The Cutting Edge Card Sorting solution.
This application is designed to help with conducting a CardSorting.

We, the development team would be happy to hear about your experiences with the application.
Please tell us about any ideas you have and any bugs you find using the forums at
http://www.sourceforge.net/projects/cardsword/

The application consists of two parts:

• CardSword:
  This is the main application, where you can create, edit and delete projects.
  This application is for the CardSorting manager, it combines all the functions he needs:
  • setup, edit, delete projects
  • add, edit, remove predefined groups (closed sorting)
  • add, edit, remove the items that will be sorted
  • create project config files for CardSword::Zorro (*.csz)
  • load result-files from CardSortings performed by users using CardSword::Zorro
  • finally create nice visual analyses and adjust these to the personal needs
• CardSword::Zorro:
  This application’s sole purpose is the sorting of a set of items into a set of (if it’s
  a closed sorting, predefined) groups. Zorro is the application that is given to the participants
  that participate in the CardSorting.
  Zorro needs a .csz-file to work, which is created by the CardSword-application.

Getting started

Installing CardSword is extremely easy:

Just unpack the zip file leaving the directory structure intact and you’re done!
There are two executable files, one for each of the two applications.

This manual will show you how to use CardSword with an example project. These are
the outlines of the example project:

The Card Sorting is about items on a website (big surprise! ). To keep the example
compact and easy to understand we will not use a very big set of items. The example
website is www.sourceforge.net. We are going to *conduct* an open card sorting.

These are our items:

• Register New Project
• Subscribe now
• Manage Subscriptions
• Advanced Search
• Direct Download
• Priority Tech Support
• Project Monitoring
• Site Docs
• Site Status
• SF.net Supporters
• Compile Farm
• Foundries
• Project Help Wanted
• New Releases
• Get Support
• My SF.net
• Software Map
• Donate to SF.net
• About SF.net

This should do to start with.

The first thing we have to do is start CardSword (CardSword.exe or CardSword.jar). Keep in mind
that CardSword is a Java application. So if you do not have a recent Java Runtime Environment (JRE) installed,
it will not work. The minimum JRE version required by CardSword is J2SE 1.4.

When you start CardSword for the first time it will generate the database it is working with. You don’t have to
do anything for this. The Database files will have the name CardSwordDB.* and will be created in the CardSword-directory.
If you like to you can also point CardSword to another database location using the Preferences dialog.

CardSword will start up with an empty project list. So the first thing we will do is
create a new project: Click the “New project”-button in the toolbar at the top. Now you can specify some
things for the project, such as name, description, the type (open or closed) and if it’s an open sorting
also the minimum and maximum number of groups the participants will have to sort into.

Click on “Save Changes” to save the project.

Basically you can just work through the tabs one after another to go through the whole
process. Therefore now click on the “Items”-tab. The “Groups”-Tab is disabled because we are doing an
open sorting in this example. The “Groups”-Tab would allow you to specify the groups the participants
have to sort the items into.

On the “Items”-tab, add all the items you want to sort.

Now the setup of the project is done. Proceed to the “Participants”-tab. You have several buttons there,
right now we are only interested in the “Generate config file…”-button. The config file that will be
generated is needed to run CardSword::Zorro, the application in which the actual sorting is done.

Generate the config file and save it where you like to. Now you could distribute this file along with the
CardSword::Zorro application (use the seperate download archive for this, it has been optimized for small
file size) to all the participants that participate in your Card Sorting project.

The process that the participants have to perform is not hard. In this example we show you through it
once. In reality you would, of course, have a lot of participants perform this task to get some results.
There is no maximum for participants btw.

So now start the
CardSwordZorro application (CardSwordZorro.exe or .jar). As this is also a Java program it also needs a
recent JRE. Upon startup, CardSword::Zorro will ask you to specify the path to the config file (*.csz)
you generated before. The config file will be loaded and the configured GUI will appear.

The items can just be dragged around from group to group. Not all items have to be sorted, the participants
are free to leave the ones they are not sure about in the item list on the left (or drag them back there).
The participants can provide a group title for each group if they want, this is not required however.
When the participants
are done with sorting, they use the “Save your Sorting” button on the top left to save their sorting in
a file. This file has to be sent back to the researcher in any way, email, floppy disc, etc.

Back in the main CardSword application, you select your project, switch to the participants tab and use
the “Import participants…” button to import the file that you (or the participant) generated with
CardSword::Zorro.
A requester will appear asking you to enter an ID or a name for the participant you’re importing. An
automatically generated ID will already be displayed, however, here you could specify your own ID for
the participant as well as add more information, i.e. “#3, male, 28 years old”. This will help you to
analyze the results one different parts of the your participant set (This functionality is not included
yet, though, right you can only run analysis on ALL participants in the project). For example you could
run the analysis on male and female users seperately and compare later (ever done a CardSorting for the
IA of a perfume store or a car manufacturer? ).

After you imported at least one participant into the project, the “Start Visualization” button will
become active. With only one participant the visualization will not tell you very much, that is why
I generated 5 sortings and imported them. This is still not enough, but at least the visualization
will show something a bit more meaningfull than with only one participant.

If you haven’t
already, click the “Start visualization” button to get some pretty graphics. The tab
that is initially open is the Dendrogram view of the cluster analysis run in the background. You can
use the slider to “cut” the graph and cluster the items. Color coding for the text and boxes around
the clusters are used to visualize the clusters. The panel on the right provides several options to
show/hide things in the dendrogram. You can change more things about the visualization in the
Preferences dialog, which I will explain further down.

The second tab
gives you a graph visualization of another kind. Items that have been sorted into a
mutual group are drawn closer to each other and are connected with thicker lines. Using the slider
on the right you can cut off all the connections between items that less than x% of the participants
have sorted into a mutual group, where x is your slider position. Play around with it. The more
participants you have, the better the results get.

Now that’s it for
the moment. One last thing: the Preferences dialog.
You can access it in the main window using the button in the toolbar at the top.

Here you can change the appearance of the visualizations:

• Change the colors that are use to color the clusters. The program cycles through the list of
  colors, so less specified colors here result in more duplicate colors.
• Change the background and the logo image.
  Choose any PNG or JPEG image here. The background will always get stretched to the size
  of the vizualisation.
  For the logo, PNGs with transparency work best. The logo will not get stretched, just placed
  “as-is” into one corner of the vizualisation.
• Change the position of the logo
• Change the fonts for the Item-Names in the dendrogram and the small texts (% in dendrogram,
  item names in JUNG vizualisation)
• Change the path to your database. You can either specify a path to a not-existing database which
  will then be generated (empty) or to a database that already exists. The database you currently
  use will not be deleted or anything if you switch to another. That way you can have
  multiple “workspaces”.