could be validated using XML Schemas and maybe some tool to check the lattice structure.
Finally the application should be able to load and save classification drafts locally, allowing users to make a valid classification over several sessions.
3.6.1 Actors
This is a description of the actor that is going to interact with the Classification-Designer. The application is intended to help the WH system administrators create and edit classifications, so no ordinary7users will ever get to see this application.
Designer - A person that has the necessary knowledge to design classifications for the system. Designers are also able to administer the contents of the WH system, for instance by uploading new classifications or by removing obsolete or faulty classifications.
3.6.2 Use Cases
Descriptions of the actions that the designers should be able to perform in the ClassificationDesigner. The use cases are illustrated on figure3.9and described in detail in the following section.
Edit Classification
Save Classification Import Classification Export Classification Create Classification Load Classification Validate Classification
Download Classification Remove Classification Upload Data Document
Authenticate Upload Classification
Remove Data Document Download Data Document
Authenticate User
Remove Classification Insert Data Document Get Data Document Insert Classification
Remove Data Document Get Classification
Effectuate Changes Set update flag
<< uses >>
ClassificationDesigner
Designer
<< uses >>
WH Web System
<< uses >>
<< uses >>
<< uses >>
<< uses >>
<< uses >>
<< uses >>
Figure 3.9: Use Cases for the Classification Designer
Use Case Descriptions
The use cases for the ClassificationDesigner can be split up into two groups - the use cases that are related to creating and altering classifications, and a group of
7The normal users that browse the WH sites.
administrative use cases. The latter being the use cases on the right side of the actor in the diagram. The WH system use cases are not described, as they are simple methods and their meaning can be deduced from the use cases that<< uses >>
them.
Create Classification Enables the designer to create new classifications. When the designer chooses to create a new classification, he is guided through a sequence of dialogs, making sure all the necessary information for the CD is in place, before he starts drawing the structure.
The drawing of the classification structure and the assignment of item refer-ences is covered in next use case - “Edit Classification”. It is possible for the designer to assign sites to classes while drawing
Edit Classification This use case covers the actual process of creating the classi-fication graph (showing the structure) and assigning item references to the class nodes.
Creation of the structure is done by inserting class nodes into a canvas and then connecting them with arrows.
After a node is created, the properties of the class, that the node represents in the classification, can be defined and item references to the data document can be inserted.
Save Classification Used to save a complete- or draft classification locally, in a specialized ClassificationDesigner format. This use case is used when devel-oping classifications. A “save” preserves everything, even the size/location of the nodes in the graph-area (the XML representation of a classification only contains structure information, no layout).
Load Classification The process of loading a classification formerly saved in the
“Save Classification” use case. This use case restores state completely.
Validate Classification An action used to check whether the loaded classifica-tion is valid, this validaclassifica-tion checks to see if all required data is entered and whether the classification graph has a legal structure (e.g. one root and no cycles). Another obvious possibility for validation is validation against an XML Schema, but this option is covered in the “Import Classification” use case.
Import Classification Import a classification stored in an XML file. This use case is utilized when the user chooses to load an existing classification from the hard-disk. When the ClassificationDesigner imports the XML, it is validated using an XML Schema.
A very simple layout algorithm is implemented, in order to “spread” out the nodes in the graph, without the layout algorithm all the nodes would be inserted in the same spot, and the designer would have to move them around manually.
Export Classification Upon completion of a classification the user chooses to ex-port it to an XML file. This file can then be uploaded to the WH system using the “Upload Classification” use case. The classification data is
vali-dated prior to export.
Upload Classification The process of uploading a classification to the WH sys-tem. The WH system relies on this action to upload valid data only, so some validation of the data must be performed prior to upload. The validation is done by validating the classification against an XML Schema.
Download Classification The process of fetching a classification in XML format from the server, and writing it out into an XML file that can be used in the
“Import Classification”.
Remove Classification The action of removing a classification from the WH server.
This action is provided in order to keep users away from the database man-agement system, thereby trying to prevent some of the mistakes that may happen in such a situation.
Upload Data Document In order to provide users a simple interface to the WH database, it is possible to upload an XML document containing the data, that the classifications in the system are indexing.
Download Data Document Used for downloading an XML document containing the data that is present in the WH system. The Data Document is used in relation to many of the editing features.
Remove Data Document Used to remove unwanted Data Documents from the server.
Authenticate This use case enables the user to actually connect to the database. It must be called prior to making changes to the WH system, or else the server will reject the changes.
Effectuate Changes Because the WH system caches views of the classifications, the system needs to know when the data is changed and the views should be updated.