What Is an Application?

In this section: Generating Samples and Tutorials

xx

An application is a platform-independent repository for a group of related components, such as procedures, Master and Access Files, data files, HTML files, PDF files, and image files. It provides an area on the server that both confers a unique identity on the application components and facilitates the sharing of components across applications in an organized manner. This construct also simplifies the process of moving a user application from one platform to another and of deploying PC-developed applications.

These components are physically grouped together on an application-by-application basis for run-time execution. This physical grouping can be within an application under a common root or a mapping to an application anywhere in the file system. The physical application or mapped name is referred to as the application name in this document. A comprehensive set of application (APP) commands are provided to control/manipulate the application components, as well as to facilitate applications that can be written and deployed to any platform.

The physical location of an application and its components is determined by a configuration parameter called approot. This parameter is set at installation time and stored in the server configuration file, edaserve.cfg. . The default value is dependent on the platform, relative to the install ID home directory, where applicable.

Application directories can be nested. A nested application directory is an application created within a higher-level application. For more information. see Managing Applications and Paths.

You can also create a home application directory for each user. Providing a user home application gives each user a place where he has full control to create, change, and run his applications. For more information, see Managing Applications and Paths.

The various operating systems on which the product runs have their own behaviors for physical files and how directories or components are referenced. In addition, some are case sensitive and some are not. For example, on Windows, the files abc and ABC refer to the same file, regardless of how it is stored on disk (aBc, for example). However, on UNIX they are all different files. Additionally, z/OS under PDS deployment makes file names uppercase when they are actually saved but they can generally also be referenced in lowercase or mixed case after they have been created. Some operating systems also allow spaces in file names and some do not. To co-exist within the various platforms on which the product runs, the product rules are to either always use APP commands or product tools such as the Reporting Server browser interface to create apps for application files (which will create them in the appropriate case), or to use lowercase names with external tools (such as mkdir myapp and vi mytest.fex), as they will save appropriately and work within the APP framework. Additionally, spaces in file names are not allowed. The use of external tools is also discouraged, as they may not act in the same way as internal tools.

Note:

• Where directories are referenced in the chart, lowercase application directories are created below the approot value. Actual application names must be lowercase and must not contain spaces. For z/OS PDS deployment, data sets are created for each component type using the approot value as the high-level qualifier.
• Some server features will not work without nested applications enabled. Among those are home application and file upload.

Two applications are provided during installation: a default application called baseapp and an application in which you can generate legacy sample files called ibisamp. In addition, when you connect to the server, a temporary directory called foccache is added as the first directory in the search path. When you want to be able to reuse data within the same browser session, you can store the data in the form of a HOLD, SAVE, or SAVB file in the foccache directory. As long as the browser session remains active, the files stored in the foccache directory can be referenced in requests.

Access to a particular application component can be explicit or implicit. Implicit access is dependent upon the search path in effect at the time of execution. The search path always includes the default application, baseapp. There is no need to explicitly declare this application.

You can change the search path from the Reporting Server browser interface, from the Data Migrator desktop interface, or from within your application code. You can also change the search path temporarily to add application names to the beginning or end of an existing search path. APP commands are described briefly in Application Commands Overview and in detail later in this chapter.

In addition to explicit APP names under APPROOT or an APP MAP command, which might look like myapp/myproc, there are also special reference names for internal locations that may be useful. They are:

While the EDAHOME and EDACONF catalog locations are internally already on a server search path, special handles such as these allow you to explicitly reference names of the form _edahome/catalog/sysapps, which is one of a number of internal catalog tables.

Note: For platforms that support the Universal Naming Convention (UNC), you can specify a network drive for approot. The UNC must be:

• Minimally one folder below the initial shared location.
• Not contain spaces. For example:

  \\mynode\myshare\accnting