Natural Secured Menus Architecture
Functional Specifications

Revision 2
September 25, 1992

University of Arkansas, Fayetteville
Computing Services
Kathryn Cantrell
W. David Wimberly

Table of Contents

Introduction

NSM Concepts

System Definition

Standards

Additional Considerations


Figures

  1. Sample menu screen
  2. Sample menu banner with parameter
  3. Sample function module banner with parameter
  4. Processing Flow of an NSM Application

Introduction


System Summary

The Natural Secured Menus Architecture (NSM-A) is both an approach for designing and developing online Natural applications and a collection of tools and facilities to assist in that endeavor. It has been designed to provide:

The NSM Architecture also provides a high level of commonality between applications so that training requirements of individuals who use or maintain multiple NSM applications is reduced significantly.


Purpose of this Document

These Functional Specifications were prepared in order to provide a clear and concise definition of the NSM Architecture. This document is intended to provide all the information necessary for an individual to develop an application using this architecture. The UAF Program Generator functional specifications should also be consulted since that facility automates the creation of most program code for an NSM application. Additionally, portions of this documentation will assist individuals in using applications that have been developed in the NSM Architecture.


Revision Identification

These specifications were first prepared and released on November 15, 1988. Revision 1 was prepared November 22, 1989 to incorporate text prepared for the presentation of the NSM Architecture at the Fifth Annual SAG University/College BIG Conference. In September of 1992, Revision 2 was prepared to include significant enhancements that were made to the NSM Architecture, prompting reference to the revised system as NSM v2 (version 2). In January 2001 this documentation was converted to HTML with some editing and updates performed at that time.

Features

Features of the NSM Architecture include:

  1. Use of a Command-Driven, Menu-Augmented (CDMA) interface so that applications will be "self-teaching, confidence-building, and task-efficient". This approach accommodates the needs of the new or infrequent user as well as the expert user. Please refer to the article "Command-Driven Application Architecture" by John B. Wheat, Third Annual University/Software AG BIG Users's Group Conference for additional information on CDMA systems.
  2. Increased productivity during system development due to a defined and tested online user interface and the existence of a code generator that accommodates most processing needs. The generated programs can be regenerated to incorporate subsequent enhancements.
  3. Implementation of a common user interface across multiple applications. NSM-A applications will look and operate similarly, thus reducing the training time of individuals who use them.
  4. Provision for command level security so that a user or group of users can be restricted from one or more commands within an application. Those restricted commands are suppressed from the user's menus.
  5. Assignment of a security level to an individual user within an application. Each application defines the actions a security level allows, and the evaluation of a user's security level is done through common code in each function module. Additionally, more restrictive security levels may be assigned to a user for selected commands through the use of command security groups.
  6. Assignment of record level restrictions (security by value) to a user or group of users. These restrictions are passed to each function module for interpretation, resulting in limited access to data.
  7. Allowance for the security administration of the application to be performed by an end user who has been designated as an application owner. This facility is provided by the NSM Maintenance System and is possible due to the data based nature of the NSM architecture.
  8. An application notice facility where messages and other general communications can be posted. A user will receive each notice only once.
  9. Provision for a temporary lock so that all users or selected groups of users are temporarily prevented from accessing an application. An explanation of the reason for the lock must be included.
  10. The ability to suspend a function in the middle of an action, go to another function and complete some other unrelated action, and then return to the original function. Three active levels are permitted.
  11. Display of a warning if modifications have been made on a screen and the user's input would result in losing those changes before they have been saved.
  12. The capability for a function to use multiple screens (up to 4). PF keys may be used to scroll between screens, or the user may specifically request a screen number.
  13. Automatic determination of the validity of each screen in a function. The user is prompted to enter data, press PF10, or go to a specific screen number to validate data there.
  14. A "Next Record" function whereby the next logical record is retrieved from the data base and displayed.
  15. A Copy action that will copy predetermined fields to effect either an add or an update to the specified key.
  16. Color and attribute settings for fields on the screen derived from user specified options (maintained as part of a user profile).
  17. Standardization of common PF Keys across all NSM applications.
  18. A central data resource of information regarding users and applications is provided. The NSM Architecture files may be used by other applications for other purposes and may be expanded to include additional data elements needed by those applications (e.g. the parameters of the user profile).
  19. Implementation of these features without adversely affecting system performance. This is made possible by the extensive use of the global data area for storing and sharing information between programs.

Background

The University of Arkansas acquired Natural and other Software AG products in the Summer of 1986. The first Natural applications developed used their own approach to menu design and relied upon the security features of Natural Security. Through the Third Annual SAG University/College BIG User's Group Conference, UAF was exposed to John Wheat of the University of Texas and their Command-Driven, Menu-Augmented (CDMA) approach to applications development. Due to the many benefits which accompanied CDMA, this technique was used in the development of two applications. There were still many drawbacks, however. The system relied upon Natural Security, which was difficult to administer, and menus included all commands even though they might not have been valid for a user. Building on the CDMA approach plus adding ideas gleaned from other SAG user group conferences, the design of the NSM Architecture was begun in late 1987. The architecture was used in an application and continued to evolve through the first half of 1988. A complete definition of the NSM architecture was developed by the Fall of 1988 and all new Natural applications were being developed in NSM.

Significant enhancements were implemented in 1992 to create version 2 of the architecture. The core architecture changes were primarily made to support the suspend function. Most other enhancements were made to the model function modules used as the base for building NSM applications. A program generator was also specially designed to use these new function module models (see the UAF Program Generator Functional Specifications).

A parallel and related development is the NSM Maintenance System. The NSM Architecture relies upon nine logical data files. The NSM Maintenance System (NSM-MS) was designed to provide the necessary maintenance facilities for these files. The specifications for the NSM-MS were prepared in the late Spring and early Summer of 1988 and development completed in early 1989. In 1992, NSM-MS was converted to use version 2 of the NSM architecture. In June 2001, the transition was made to position based desk assignments which necessitated changes to NSM-MS and to some core components of the architecture.

Additional information regarding the NSM-MS application is available in the Functional Specifications for that system.


NSM Concepts

There are several key concepts that should be understood before reading further about the NSM Architecture. Each is identified and discussed below.

Commands

Commands provide the transportation for traveling to various destinations within an application while menus serve as road maps. More precise definitions for these components are:

  1. A command is a short name used to identify an online function or menu
  2. A menu is a list of functions or sub-menus available within an application
  3. An application is a collection of functions designed to address a general need

Control is passed from one function or menu to any other by entering the desired command. A command will usually require only one screen, however some complex functions may require several screens or windows.

When a value is entered in the Command field, it is validated and results in an error message, display of the menu requested, or transfer of control to the requested function associated with the command. If a command is entered and PF2 is pressed, the current function is suspended while the new function is executing. PF3 returns the user to the original function at the point of suspension.


Menus

A menu contains an area where a command can be entered, an area where key field values may be specified, and a list of valid and related commands. The list of commands is accompanied by descriptions and identification of any required key fields for the commands. Commands may be selected in any of three ways:

  1. The desired command may be entered into the field Command.
  2. A non-blank value may be entered into the field that precedes the desired CMD within the list in order to mark it for selection.
  3. The cursor may be positioned to the line which contains the desired CMD and the Enter key pressed.

Values may also be entered into any of the key fields that appear in the top portion of the screen (the banner). Key field values which are entered on the menu will be passed to other programs and thus speed up the operation of the system by allowing the command and key field specification to be entered in one step. The Required key fields column on the menu identifies the key fields which are used by each command.

See Figure 1 for a sample menu screen from the NSM Maintenance System.

Figure 1. Sample menu screen


Enter, mark or position cursor to desired command                             
 NSOMENU 1 TEST       nsm maintenance system Main Menu - MM      01/04/01 09:34
 Command:      Action: V  Desk:           Appl:           Cmd Sec Grp:         
 Val Sec Grp:           Parm:                                                  
-------------------------------------------------------------------------------
   CMD  Command description                      Required key fields           
   ---- ---------------------------------------- ------------------------------
 _ LAAC List All Available Commands              Command ID                    
 _ MHAC Menu Hierarchy of Available Commands     --none--                      
 _ DAM  Desk Administrator's Menu                                              
 _ ISM  Initialization and Setup Menu                                          
 _ AAM  Application Administrator's Menu                                       
 _ HELP nsm-ms system level HELP                                               
 _ CDCD Change desk/Display Current Desk info                                  
 _ LOG  LOGon to another natural application                                   
 _ FIN  FINish your session (sign off natural)                                 
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
                                                                               
Enter-PF1---PF2---PF3---PF4---PF5---PF6---PF7---PF8---PF9---PF10--PF11--PF12---
      Help        Quit                                                   

Key Fields

Most online functions require one or more key field values to be specified in order to define what is to be done. These values usually define what record or records are to be operated on and/or what operations are to be performed. The required keys are normally output modifiable fields in the top portion of the screen, the area above the dashed line known as the banner area. Once a value is entered in a key field, it is retained until another value is specified. In this manner, several commands may be accessed to operate on the same entity (a student) without re entry of the key (student ID).

The required key fields for a function define what is to be done. These fields, and their corresponding tags, appear intensified (or in a different color) within the banner area so they can be easily identified for each function. The entry of new key values is treated as a request to initiate a new operation within the current function and/or to perform that operation on new data. If a data base update was intended for the previous key values, it would not be completed unless PF10 had been pressed. However, a warning message is issued if modifications have been made but not saved and new keys and/or a new command is issued. When the key values specified are not valid, an appropriate error message is displayed and re entry of the erroneous field requested. In this manner, the current function module is used to perform various operations on records until another command is requested by entering a Command or exiting from the function module via PF3, Quit.

The banner area may contain one or two lines for the key fields. When possible, the format and content of the banner area should be consistent on all screens, i.e. the same key fields in the same place with the same tag. This is often not possible since there is a limited amount of room in the banner and some applications will utilize many key fields for various functions. A technique to accommodate this situation is to place the most common fields in the banner and leave space in one area for unique fields needed by only one function module. This method, however, prevents those unique key fields from being passed from the menu or from other functions. This quandary is addressed by providing an alphanumeric parameter field in the banner, defined as input only. Functions which utilize unique key fields then determine if there is a value in the parameter field, edit it for format consistency, and move it to the unique key field. When calling these functions from other function modules or a menu, the desired key or keys is then entered into the parameter. See Figure 2 for the banner area from the menu of an application using this technique and Figure 3 for the banner as it appears in a function module.

Figure 2. Sample menu banner with parameter


 DAMMENU                  TaBLe maintenance - TBL                02/17/89 16:52
Command: acct  Act: v ISN:            SSN:            Date: 02/17/89 Dept:
                             Parm: aaaa          DcReg:          Pd#:
-------------------------------------------------------------------------------

Figure 3. Sample function module banner with parameter


 DAMACCT             ACCount Table maintenance - ACCT            02/17/89 16:53
Command: ____  Act: V ISN:            SSN:            Date: 02/17/89 Dept:
Account Cd: AAAA             Parm:               DcReg:          Pd#:
-------------------------------------------------------------------------------



Action Key

Traditional file maintenance functions allow for record addition, deletion, update, or view. These actions can be implemented with four function modules, each designed to perform one of these actions. However, the screens needed for each action are very similar and much of the processing is identical. Each of these actions and more can be implemented in one function module as long as there is a way to determine what action is appropriate. This is accomplished by including a key ACTION field in the banner. Although special purpose action codes may be needed to address unique requirements, the following are standard NSM values for the action field:
Action Value Description
A Add
U Update
V View (read only)
D Delete
C Copy (effect either an Add or an Update)
This ACTION field allows the same program and map to be used to perform adds, updates, views, deletes, or copies, thus minimizing the number of programs to be developed and maintained. NSM security features allow all users to access a multi-purpose function module, while an individual's security level can determine which actions are appropriate. The modules of the NSM-MS are examples of this technique.

Validations

All input data is checked and edited for accuracy (to a degree) prior to it being accepted and stored on the data base. This is accomplished through the execution of validation rules. When the user enters or modifies data on a screen, the execution of these rules may result in an error message. Error messages are displayed within the Natural message line and the cursor is placed on the field associated with the error. The user is not allowed to save the changes to the data base until the error is corrected.

In multiple screen functions, it may be required that data be entered and validated on several screens. In this situation, PF10 will not be labeled "Save" until valid entries have been made on all screens other than the current screen. The system will also prompt the user, identifying the screen number, when valid input data is required on a different screen (given the current screen passed all edit checks).


Warnings

Warnings are messages displayed to the user for information purposes. They appear much like error messages, but always start with the the text "WARNING:". Unlike error messages, however, warnings are issued only once for the same screen, key fields and cause. One screen may issue many different warning messages for different reasons (causes).

An example of a situation where a warning would be issued is when a user makes changes to the data on a screen and then requests another command. Because PF10 was not pressed, a warning message is displayed that informs the user that the changes that have been made will not be saved. The user then has the option of pressing ENTER again to execute the command (and lose those changes), press PF2 to suspend to the requested function, or press PF10 (which will save the changes and then execute the command requested). If the user opts to suspend the current function, changes will not be lost. However, when the user returns to the suspended function, any other action that could result in the loss of modifications to the current record (e.g. PF3) will NOT result in a warning message. The warning for that reason for the keys being operated on has already been issued. Once a warning has been given, the user is allowed to proceed at his or her on discretion.


PF Keys

All valid PF keys are named and the names displayed on the last two lines of the screen. They are displayed in either Natural format or a PC style format based upon the user's setting in the User Profile. PF keys that are not valid in a certain circumstance will have their names removed so that they do not appear as valid options. The error message "PFxx is unassigned, valid PF Keys are displayed on the screen" is displayed whenever an inappropriate PF key is entered. The standard PF key usage is:

Key & Name

Usage

Enter

If new key field values or a new command have been entered, initiate processing for the new keys or the new command issued. Otherwise, validate any entries made in the body of the screen and prompt for the next activity.

PF1 (Help)

Initiate help processing. This will be field help if the cursor is on a field defined with a help routine; otherwise, screen level help will be provided.

PF2 (Suspd)

Suspend the current program and transfer control temporarily to the requested function module. Entry of a command is required. Suspend is not valid from a menu.

PF3 (Quit)

Quit the current operation and back up one level in the processing hierarchy:

  • If within a window, return to the primary screen.
  • If within a function module and at level one (no programs are suspended), return to the menu which contains that function module.
  • If within a function module and at a level greater than one, return to the program suspended at the next lowest level.
  • If within a menu, return to the menu that contains that menu command, or if on the main menu, return to the LOGON application. If the current application is the user's default application, sign off Natural.
  • If within the application selection screen of the LOGON application, sign off of Natural.

PF4 (DCode)

Decode selected coded values on the screen by performing lookups and displaying associated names, descriptions, or other information. This key may also be used to explode a selected entry by displaying within a window more detailed information than was available on the original screen.

PF5 (RStrt)

Restart the operation with the keys specified in the banner area (disregard any changes or entries made within the body of the screen). On list functions, any previously marked entries for selection are cleared.

PF7 (Back, PrevS)

Page backward to values or previous screens displayed within the same function module.

PF8 (NextS, NextR, NextT, Q/Nxt, Forwd)

PF8 always moves the user forward to access the next set of data, although the key is labeled and the function is performed in a different manner dependent upon the circumstances.

  • If on a multi-screen function, this key is labeled "NextS" and pages forward to the next screen.
  • On the last screen of a function, or when there is a command or key error, it is labeled "NextR" and will retrieve the next logical record for display.
  • Additionally, within a function that is operating at level 2 or 3 and which was accessed from a suspended list, this key is labeled "Q/Nxt" for quit this function and retrieve the next selection. If no other entries are marked, the effect is a quit to the suspended list.
  • At other times, PF8 is labeled "Forwd" and pages forward to display the next set of records.

PF10 (Save, Sav/Q)

PF10 is the commit key. No changes are effected or permanent action taken within an NSM application unless PF10 is pressed. It will save onto the data base any changes that have been made or data that has been entered. It may also commit or finalize the action being performed, such as submission a batch job for processing based upon the specified parameters. Within a function module operating at level 2 or 3, the key is labeled "Sav/Q" to save the changes and then quit that function and return to the suspended function.

PF11 (Options)

This key is provided for the selection of options that may be available in individual function modules, or may indicate a specific option when further selection is not required.

PF12 (Print, Email or Flip)

On lists where a print function has been associated with the listed items and the user has selected (by default via the user profile or use of the RODS command) a Report Output Destination ID, this key will generate a print (report) of the selected items. Similarly, it may be used to generate an e-mail of the selected information. On function modules where more than 12 PF keys are defined, PF12 is flip and will flip the display of PF keys to show PF13 thru PF24.

PF24(Flip)

This key flips the display of PF keys to show PF1 thru PF12.


Update Confirmation

Data base updates are only performed in response to an explicit request. Pressing the ENTER key will cause all inputs to be validated, but it does not result in any data base change. Any input errors will be reported and the cursor placed on the invalid field. When all inputs are valid, a prompt will be displayed indicating that the entries should be saved by pressing PF10. Pressing PF10 serves as the explicit request to perform an update. Following all updates, a confirmation message is displayed indicating the action that has been performed. This confirmation will override the line of the screen containing the program name, program level, and function title (normally line 2 of the screen).

If a user is confident of his input and speed is essential, input data may be entered, new keys and/or another command specified, and PF10 pressed in the same step. If there are input errors the result will be the same as if ENTER were pressed; otherwise the update will be performed, the confirmation displayed, and the new keys and/or new command processed. If the only error is with a new key specification or command id, the update is performed, the confirmation displayed, and the appropriate error message displayed. This technique provides the power user with an alternative data entry method that supports fast and efficient operation in the production environment.


List Functions

There is a special class of NSM functions designed to list information from the data base. These are inquiry only functions (no update capability) which allow browsing and selecting data online. There are several variations of these list functions, but they all share the following features and characteristics.

  1. Key fields are required to be specified to define what data is to be listed. The key fields may either define specific data to be listed (all employees in BU HIST) or, more commonly, define a starting point from which data is listed sequentially (employees in BU Hist with names starting with Williams).
  2. PF8 (Forwd) is permitted to page forward through the data sequentially and PF5 (Rstrt) will restart the list. Most lists also support PF7 (Back) to page backward.
  3. The screen will display a count of the number of entries displayed from the starting key value, and in some cases will indicate how many total entries exist. This will appear as:
    Employees 11 thru 20 of 39 displayed
    
    The starting keys used to retrieve the data are also displayed immediately below the banner.
  4. Entries from the list may be selected so that they might be operated upon further. The selection may be performed by marking (entering a non-blank character in the input field preceding each entry) or by merely placing the cursor on the line containing an entry and pressing ENTER. The effect of selecting an entry is that the key fields from the entry (those fields which are intensified) will be placed into the banner keys so that they will become the operands of other functions. For example, from the list of employees in a BU, the entry for employee 101202 could be marked. That Emp ID would be placed into the banner key and when a function is accessed that requires an Employee ID, information for that employee would be retrieved and displayed.
  5. The ability to mark entries is very powerful when combined with the suspend feature of NSM. Multiple entries may be marked, a new command entered (one that operates on the keys that will be selected from the list) and PF2 pressed. The effect is that the keys for the marked entries are selected and the new command invoked. Within that command, PF8 is labeled "Q/Nxt" so that it will quit that function and return to the list so that the next marked entry can be selected. This represents a very fast and efficient way to process, in detail, selected entries from a list. To elaborate on the above example, selected employees could be marked from the list and the user could request to suspend to a function enter leave data, enter hourly time, or view detail position information. This data could then be updated or viewed sequetially those selected employees by pressing PF8.

Suspend

The suspend feature allows a user to access a second or third function while still retaining in a suspended state the activity within the original function, including any data entered prior to the suspend request. The ability to suspend is most useful in the following situations:

The suspend feature is invoked by entering the command and desired keys for the new function and then pressing PF2. The current function will be suspended temporarily, and the new function module will be invoked. PF3, PF10 and sometimes PF8 will return the user to the suspended function and restore the key fields to the values which were previously in effect. A warning message is provided whenever a user's actions would cause a suspended function to be lost (e.g. enter a command without pressing PF2). Three active levels are permitted, but suspending to or from a menu is not allowed. The level at which you are currently operating is displayed on the first line of the screen as the second element (just after the program name).


Marking Entries for Selection

List functions are designed to allow selection of one or more entries from the lists provided. Values selected may be placed in the banner key fields so that they may be used to specify additional processing by other commands or the selection may be used to "decode" or explode the entry by displaying additional information about it within a pop-up window. The selection is made by marking an entry by entering a non-blank character in the field that precedes each entry of the list. If only one entry is to be selected, the cursor may simply be placed on the line of that entry. PF4 can then be pressed to decode the selected entries or another command entered to operate on the key fields selected.

Selection of listed entries is very helpful in conjunction with the Suspend feature of the NSM Architecture. Another command is entered, the desired entries are marked, and PF2 is pressed to effect the suspend function. In this case, the first marked entry is placed in the banner for operation by the requested function. PF8 is then used to retrieve the next marked entry until either PF3 (Quit) is pressed or there are no further selections. At that time, the suspended function is again presented.


Screen Selection

On multiple screen functions, there are two ways a user may scroll through the screens. PF7 and PF8 are used to page back or page forward one screen, respectively. Secondly, an explicit screen number can be entered in the field beside the label "Screen," the field directly below the banner that also displays the current screen.


De-Code

A great deal of data stored in computer systems is coded to reduce the time required for entry and the space required for storage. Often times there exist tables where coded values are stored along with descriptive text that identify the meaning of the codes. These tables are used to validate the data upon entry and to translate codes to descriptive and meaningful names on output.

In an on-line environment, system efficiency is critical and access to tables consumes scarce and critical data base resources. For this reason, descriptions associated with coded values are sometimes not automatically accessed and displayed on-line. This information is, however, available upon request. The method for requesting the de-coding of coded values is by pressing PF4 whenever this key is labeled "DCode".

In addition, the de-code function may be programmed to do more than just look-up descriptions for codes that appear on the screen. In some situations it is also used to explode data -- to access much more complete and detail information and display it within a window. This is often the case when there is a list of summary information displayed. A specific entry may be exploded by placing the cursor on the desired entry prior to pressing PF4.


NSM Field Types

There are four special types of data fields which are used on NSM on-line screens. These are:

Field Type

Description

Modifiable Default

Fields that are modifiable but which are not being emphasized and are therefore displayed with default intensity.

Modifiable Intensified

Fields that are modifiable and are displayed intensified (bright) for emphasis.

Conditionally Protected

Fields for which update would normally be allowed but due to certain conditions have been conditionally protected and may not be modified (e.g. a view or delete action is being performed or an error exist in the key specification).

Previous Value

Fields that show a previous data value prior to an update transaction being entered (as opposed to the proposed value which represents the change being effected by the transaction). Previous value fields are used in TARGET functions.

These field types have associated attributes and colors that can be customized by the user via the user profile. For instance, a user may prefer that all the modifiable default fields be shown in reverse video (i.e. the attribute) and in the color green while modifiable intensified fields are displayed in reverse video turquoise. If your terminal supports reverse video, that attribute is suggested for previous value fields since it will allow detection of a prior blank value.


Desk Definition

A desk, identified by a Desk ID, represents an individual's job, specifically those tasks related to the use of computerized administrative systems. Desks are associated with and defined within budgetary units. The department heads, or individuals they choose to designate (refered to as "Departmental Desk Adminstrators"), are responsible for the creation and definition of the specific desks within their department. The definition of a desk includes an optional description of the functions or responsibilities of the desk, as defined and to the extent desired by the department. The departmental desk administrators then associate employees within their budgetary units to the desks they created.

The desk concept permits several individuals to be associated with the same desk when they perform the same job (e.g., registration data entry operators). However, the norm is one desk is associated with one individual since most jobs are unique. In addition, one employee is typically associated with only one desk. That desk is the employee's primary desk. An interim desk may be assigned in exceptional cases when an individual is asked to perform two jobs on a temporary basis. A user who is assigned to two desks is required to explicitly identify which desk he will be working at any one time. It is recommended that this dual desk feature only be used in situations where there are true interim assignments.

The association of an employee with a desk is based upon the type of employee and the type of user ID. For appointed employees with one primary user ID this is accomplished by assigning the desk to the PSB position the individual is filling or hired into. Some appointed employees have secondary or temporary user IDs for the purpose of loaning those IDs out to temporary employees on a short term basis. Desks for these IDs are assigned directly to the User IDs. Similarly, desks for hourly employees and non-University employees that require access are also assigned directly to their User IDs.


Desk Function

Application access privileges and required electronic transaction review and approval are determined by the employee's desk assignment. A discussion of these two separate uses of desks follow:

Application Access

Desk-based security allows the access rights to all necessary applications to remain the same, regardless of who is filling a position (assigned a desk). The department head requests access to an application for a desk. The application owner establishes and maintains the access rights for a desk to an application. The department head can assign as many employees as desired to work that desk. When a user signs in with his individualized ID, the system automatically makes the association of that user/employee with the established desk assignment. The access rights and privileges of that user are those established for that desk.

Transaction Approval

For the purpose of routing electronic transactions for approval, a desk may be established as a review point for those transactions. The necessary routing definitions for electronic transactions are determined by the University administration, and are rather complex and voluminous in nature. The use of a desk ID to define a review point for a transaction permits the reassignment of personnel to be accommodated without requiring the redefinition of the routing information.


User ID and its Status

A user of online Natural applications must be issued a CICS ID. This User ID is assigned to the individual employee and should not be shared or assigned to someone else. When your department hires a new University employee who requires access to Natural applications, a CICS ID must be requested from Computing Services. This ID provides an audit trail of the activity performed, such as who updated a record, who initiated a transaction, and who reviewed a transaction.

A User ID can have one of three possible codes for its status: active, inactive, or temporary. Active status indicates the user is an active employee on the University's payroll system. (There is an exception for certain individuals who are not employees but have been granted access to University administrative systems.) Inactive status is set for a user who is no longer active on payroll and is therefore no longer authorized to access computer applications. Of course, desks cannot be assigned to inactive users. Temporary IDs are special types of IDs that are designed to fill the need of temporary workers. Temporary IDs are assigned to be the responsiblity of a supervisor in the department. That supervisor then distributes the ID and its password to the temporary worker. When the temporary worker leaves, the supervisor changes the password so that security is maintained. The supervisor is responsible for the actions performed by users of temporary IDs.

Active and temporary User IDs are associated with a budgetary unit based upon their type of employment, either appointed or hourly. Appointed employees are hired into a position which has been allocated to a budgetary unit. Hourly employees must have an active wage rate established for the budgetary unit in which they work. This budgetary unit association plays a significant role in the desk assignment process.


Desk Assignments

A desk is created for and designated as belonging to a specific budgetary unit. It may then be assigned to either a PSB (Personal Services/Budget) position (for appointed employees) or to a user ID (for hourly employees, temporary User IDs, or non-employees) which is associated with the same budgetary unit. These desk assignments are time stamped -- each has a designated beginning point in time and ending point in time (12/31/2099 designates an undefined ending period). Future assignments can be entered or changed, while historical assignments are frozen in order to provide an historical record of all desk assignments. The user's desk for a session is then determined by the system at signon using these desk assignments. This determination is based upon the status of the User ID according to the following steps.

Active users
  1. If the user is not on the University's Employee file, the desk ID (if any) associated with the user is assigned. If there is no desk for the user, the user is only permitted default access.
  2. The system determines if an employee is filling (appointed in) a PSB position on that date.
  3. If appointed, the desk ID assigned to the position at that point in time is assigned to the user. If there is also an interim desk assignment, the user is given the choice of which desk to be assigned. If there is no desk for the position, the user is only permitted default access.
  4. If not appointed, the desk ID associated with the user is assigned as long as the employee has an active hourly wage rate in the BU of the Desk. Otherwise, no desk is assigned and the user is only permitted default access.
Temporary users
  1. If the user is not on the University's Employee file, no desk ID is assigned and only default access is permitted.
  2. The system determines if an employee is filling (appointed in) a PSB position on that date.
  3. If appointed, the BU of the appointment (position) is checked against the BU associated with the desk designated for the user. It these are the same, that desk is assigned to the user.
  4. If not appointed or the BU of the appointment does not match that of the desk, no desk is assigned and the user is only permitted default access.
Inactive users
Normally, inactive user IDs will not be permitted to signon to the administrative system since the user ID should have been inactivated within the operating system. If a user flagged as inactive does reach Natural, it will not be assigned a desk and will only be permitted default access.

This system has the primary advantage that for appointed employees, no maintenance is required when an employee leaves and the position is later filled by another individual (other than obtaining a CICS User ID if necessary). Further, access is automatically discontinued upon termination or transfer within the University, either for an appointed employee no longer assigned to a position or an hourly employee whose wage rate is inactivated.


System Definition


Overview

The Natural Secured Menus (NSM) Architecture is a standardized approach for designing and developing online applications in Natural. The NSM Architecture includes several Natural subprograms and template objects which perform functions common to all applications. Each of these components is presented later in this section. A program generator is also available so that application function modules can be quickly and efficiently created using the most up-to-date NSM standard models. (See the UAF Program Generator Functional Specifications for more information).

The NSM Architecture is data based and requires the maintenance of information pertaining to applications, menus, users, desks, and security restrictions. It is intended that the responsibility for maintenance of this data be shared between the central data processing organization, the designated application administrators (also referred to as the application owners), and the departmental desk administrators. The NSM Maintenance System has been designed to provide this division of responsibility.


System Flow

The general flow of control through all NSM v2 Architecture systems is the same:

  1. The user signs onto Natural, is assigned a Desk ID, and chooses an application to use.
  2. The application entry module is executed to validate the user's access, establish the necessary control parameters, and display any application notices the user has not yet received.
  3. The menu program is executed to display menus.
  4. The subprogram 'NSNVCSV' is called to validate any commands issued.
  5. Any valid function module is executed until another command is issued or PF3 is pressed. Commands may be for other function modules or for menus. When a function module command is entered and PF2 is pressed, the current function is suspended by the use of a FETCH RETURN to the new function. A PF3 will return to any suspended function or, if none, invoke the menu program and display the menu containing the current function module command.

This process is depicted graphically in Figure 4.

Figure 4. Processing Flow of an NSM Application

Processing flow of an NSM Application


Data Requirements

The NSM Architecture utilizes nine logical files. These files are defined within the Predict Data Dictionary. Detail information from the data dictionary has been made available in the file and data element definition and the file and element summary. A summary description of these files follows.

File

Description

APPLICATION

Contains information related to the online applications: application ID, application name, main menu command, notices, valid actions and security levels, application owners, etc.

APPLICATION-DESK

Identifies the desks that have explicit access to the application and the type of access: security level, command security group, value security group, and job classes permitted.

APPLICATION-USER

Identifies the users that have accessed an application and the last date of access.

COMMAND-SECURITY

Identifies the command security groups that have been defined for an application and their associated command restrictions. All desks defined to access an application must be assigned to a command security group.

DESK

Contains information related to desks which are created by departments to define the roles or computer usage needs of their personnel. Information contained here includes desk ID, budgetary unit, and short and long descriptions.

DESK-ASSIGNMENT

Contains time stamped (beginning and ending effective date and time) desk assignments for positions or user IDs.

MENU-COMMAND

Contains the information for each menu within an application: the menu command and description, and the commands, descriptions and required fields for each command on the menu.

USER

Contains information related to users of the online systems: user ID, name, SSN, preferred field attributes and colors, etc.

VALUE-SECURITY

Identifies the value security groups, if any, that have been defined for an application and their associated value restrictions. The use of value security is an optional feature. If used, all desks established with access to the application must be assigned to a value security group.


Template Components

The NSM Architecture provides several template or model components for building a new application, all of which can be found on the SYSTEM library. A template global data area and the map used for menus are provided and must be installed in the new application's library, renamed and then modified for the specific application. The 'xx' within the names of these components is replaced with a standard application prefix used at the University of Arkansas. All maps and local data areas for the individual functions must be built by the programmer, but guidelines and models for these are also provided. This process is described further under the heading "How To".

Global Data Area (xxGLOBAL)

The NSM Architecture depends heavily upon the global data area as the means of maintaining profile information on the user and the application, as an area to hold menu information in order to minimize data base accesses, and as the means of passing key fields from one menu or function module to other function modules. The required global data fields for the NSM Architecture are in the data area xxGLOBAL. This data area should be copied for each application and modified to include the key fields and other shared values used by the application.

Menu Map (xxMMENU)

This is the only map required to display an NSM application's menus, regardless of the number of menus defined. Because most of the menu display and processing is common for all NSM Architecture applications, a model menu map is provided. This map is xxMMENU and it includes a processing rule that determines the command selected by the user. The banner of the map should be customized to contain the key fields that are desired. Generally, there should be no editing or validation of these key fields on the menu map; that is the responsibility of each function module. The menu provides these fields as a convenience to the expert user, so that the necessary keys may be provided in advance and an extra screen and terminal I/O avoided within the function module.

Templates for Standard Models

All models developed for the NSM architecture have template local data areas and maps stored on SYSTEM. Please see the UAF Program Generator documentation for a current list of models and their associated template local data areas and maps. Refer to the section "How To" for detailed instructions on using the maps.

Generated Programs

The program generator is used to create the entry program, the menu program, optionally a help program, the error program, and the function modules. Several function module model programs have been developed. Up-to-date information on these models can be found in the UAF Program Generator Functional Specifications.

Entry Module (xxOENTRY)

Each application using the NSM Architecture will have one entry module or startup program. This program is automatically executed whenever a user signs onto the application (a Natural Security definition is required). This program must be unique for each application since its primary function is to initialize the global data area elements which are used by the NSM Architecture. The primary functions performed by the entry module are:

The entry module is easily generated using the Natural ISPF edit macro PGZENTRY.

Menu Program (xxOMENU)

The menu program is responsible for displaying each requested menu. When a command is requested from a menu, the subprogram 'NSNVCSV' is called to validate it, and the menu program will fetch the valid function module, display the appropriate menu, or provide an error message. PGZMENU is the Natural ISPF edit macro that generates the menu program.

Application Level HELP Program (xxOHELP)

This program supplies application level HELP. A list of up to 14 help topics is displayed, and the user can mark any number of topics desired. The subprogram UANSHSS is the routine called to display the help information from the appropriate Predict conceptual field. After a help text is shown, the user is brought back to the HELP screen where he can then press ENTER again to see the next marked help selection. PGZHELP generates this program.

Error Processing (xxOERROR)

An error program provides a way of trapping, interpreting, and recovering from online errors without having to code this logic into each online program. The error program name is assigned to the system variable *ERROR-TA by the entry program. The error routine performs the following functions: If a program includes its own ON ERROR statement, that logic will be invoked instead of calling xxOERROR. A program can even include an ON ERROR, check for a specific error condition, and if the error is not one it expected it can still call xxOERROR. However, be aware that if a program invokes subprograms or subroutines, you will not be able to determine in which module the error occurred. This is not a problem with an error program since Natural stacks the name of the module where the error occurred. PGZERROR generates this program.

Subprograms

The NSM Architecture utilizes several subprograms in order to minimize the amount of code that must be maintained. These subprograms reside on the SYSTEM library where they can be shared by all applications. This is also very efficient since Natural is reentrant, i.e. all users of any NSM application share the same copy of the subprogram in the buffer pool (as of Natural V2.1.4).

NSNVULP2

This routine, modified from the original 'NSNVULP' for NSM version 2, validates the user's access and loads necessary parameters. It is called by the entry module when a user first signs onto an application. It is more involved than this title indicates since it performs all of the following:

NSNREADU

The subprogram NSNREADU is a utility program which takes as input a user id, looks up the id on the USER file, and returns the user's name in first-middle-last name sequence.

NSNRDUD

This subprogram performs the same functions as NSNREADU and also returns the user's budgetary unit.

NSNVCSV

The subprogram 'NSNVCSV' is used to validate a command and set command dependent variables. It is called by the menu program and all function modules and is thus a critical part of all NSM v2 applications. This routine determines if the command id issued is a menu or a function module, whether or not the user is allowed to execute the command, and, if valid, the user's security level within the command. Appropriate variables are set for each case so that the calling program can determine if a menu should be displayed, a function module fetched, or an error message displayed.

Local Data Areas

All of the standard NSM v2 programs (entry, menu, and the standard models) utilize local data areas in order to minimize the amount of code to be maintained uniquely within each application.

NSLENTRY

This local data area includes the common field definitions needed by all entry programs.

NSLMENU

This local data area includes the common field definitions needed by all menu programs.

NSLHELP

This local data area includes all the necessary local field definitions required by the HELP program.

Copy Code

There are a few areas where copy code has been used in order to minimize the code that must be maintained in application specific programs.

NSCVCSV

The copy code NSCVCSV contains the call to the subprogram 'NSNVCSV' and the variables required for it.

NSCVULP2

The copy code NSCVULP2 contains the call to the subprogram 'NSNVULP2' and its required variables.

UACPFKEY

The copy code UACPFKEY should be included in each function module prior to customizing the PF Key assignments. It will set all PF, PA and Clear keys to be program sensitive, set all PF Key names to blank, and, on the test system only, set the Clear key to the terminal control command '%%.'

Help Routines

There are two standard help routines defined for NSM applications. For more information, please see "UAF's Standard Natural On-Line Help Routines."

NSHSHM

This routine is specified as the screen help routine for the first version of NSM. It gives the user the option of receiving screen level help for the map being executed or definitions of commands, menus and key fields. The application's specific screen level help is read from Predict's extended description for that map (the map name is passed to the help routine as a parameter).

NSHSHM2

This help routine should be specified as the screen help routine on the "Map Settings" screen of the map editor for the NSM Version 2 primary function module maps. The user can select help for the specific screen of the function being executed or other general NSM topics associated with function modules. Screen level help is obtained from Predict's extended description for the appropriate map.

NSHMENU

This routine should be specified as the screen help routine on the "Map Settings" screen of the map editor for the menu map (xxMMENU). It provides options for receiving help information on general NSM concepts such as menus, commands, key fields, standard PF Key usage, definition and features of the NSM Architecture, and reporting of problems and suggestions.

UA Programs

Programs may be stored on the SYSTEM library to perform common functions for all applications. These programs may be included as commands on the application's main menu. Two of these programs that are applicable to all applications are identified here.

Note: The Type on the MENU-COMMAND file for these commands must be 'U'.

UAOFIN

This program is a simple method of implementing the terminate or finish function. FIN is the command that must be defined on a menu in order to invoke this routine. The program merely executes a TERMINATE statement.

UAOLOG

This program provides a window so that the user can specify another application to access (logon to). It also provides access, via PF6, to the user profile maintenance function.

Standards

There are several standards that should be followed when developing NSM-A applications. These standards are automatically incorporated when the program generator is used with the standard NSM modules.

Program Name Format

All command IDs are limited to one to four characters within the NSM Architecture. The subprogram 'NSNVCSV' constructs the program name to be fetched (if it is valid for the user) from the application prefix and the command ID. The resulting program is then fetched by the calling program. The format of program names is xxOcccc where:

xx

The application prefix as defined on the APPLICATION file.

O

The UAF standard for online programs.

cccc

The command ID.

All primary function module program names must conform to this format. A function module may fetch or call another program with a different format name. When this is necessary, dashes can be used as fillers and the eighth character included to identify related modules. For example NSOU may fetch NSOU---1 to perform some associated processing.

Map Name Format

Map names are derived directly from the command id for the function module. The format of map names is xxMcccc where:

xx

The application prefix as defined on the APPLICATION file.

M

The UAF standard for map names.

cccc

The command ID.

If the function contains more than one screen, it is required that the map names be in the format xxMccccn, where n is the number of the screen. If the command id is less than four characters, the remaining characters must be filled in with dashes ('-'). For example, for the two-screen function associated with command U (User maintenance), the map names are: NSMU---1 and NSMU---2.

Note: Maps used for decode also can follow the format xxMccccD, whereby the remaining positions of a command id (if less than 4) are filled in with dashes (e.g. NSMLAO-D is the decode map for the NSM-MS list function LAO).


Key Field Display

All key fields that are required by a function module should appear as output fields in the body of the screen (below the banner). The same fields will also be in the banner; however, they indicate what keys will be processed next in the banner. Below the banner, the keys indicate the record currently being processed, or, as is the case with table and list modules, the starting record. This should prevent any confusion regarding the association of the data displayed with the key fields in the banner. This situation occurs when a record is being updated and new keys are specified in the same operation.

Key Field Validation

The key fields for a function module should be validated prior to the execution of the input statement. The prompt displayed with the input statement should then indicate what the user should do next, based on the validity of those keys. Through the use of control variables, the fields that are normally modifiable within the body of the screen can be protected if the keys provided are invalid or the action requested necessitates they be protected from entry. See the section "Single Screen and Multi-Screen Function Modules" for additional recommendations regarding function module design.

Case Sensitivity

Each application should establish its treatment of lower case data at the time the user logs on. The entry module executes a SET CONTROL 'L' or SET CONTROL 'U' dependent upon whether the application needs to accept mixed case input.

Confirmation Messages and Prompts

The Natural message line should always contain a prompt to indicate what the user should do next. In the case of an edit error, the notification of the error and requested correction is an appropriate prompt. In addition to a prompt, a confirmation of any action just completed is to be displayed on the first line of the map (not the message line, which may be positioned at the top or bottom of the screen). When present, this confirmation overlays the the program name, level, and command title. A confirmation can indicate what records have just been added or updated or that the current operation has just been restarted. The confirmation and prompt fields are assigned values conditionally based upon the program logic. Utilizing this technique, one operation can be confirmed following a PF10 at the same time a prompt is generated as a result of new key fields and/or a new command being entered to initiate the next operation. Confirmations and prompts should utilize mixed case for readability and should not contain periods.

Field Highlighting

The labels and the fields for the required keys in the banner and the key fields displayed directly below the banner should be displayed intensified. All other tags and output-only fields should be default intensity. Fields that are conditionally protected should be default intensity when protected and intensified otherwise.

These general characteristics may be customized by a user in his user profile by selecting an attribute and color for fields characterized as follows and retained in the control variable indicated:

Description Control Variable
Modifiable default intensity #MD-FA
Modifiable intensified #MI-FA
Conditionally protected #CP-FA
Previous value #PV-FA
The user selected attribute and color designations are assigned to global control variables by the entry program and may then be used in maps. The standard use of these control variables in the banner is as follows:

The required banner key fields use the control variable #MI-FA, while the remaining key fields use #MD-FA. On the menu, only the 'Command' field should be intensified (#MI-FA).

There are other control variables used in the standard NSM models and are required on elements within the body of the map. See the Program Generator Functional Specifications for further information.


System Level Help

System level help can be provided for each application by developing a HELP function module. This help facility should include: The UAF Program Generator facility provides the method of generating this HELP function module. xxOHELP is described in the section "Generated Programs", and further information on the HELP model can be found in the UAF Program Generator Functional Specifications.

Single Screen and Multi-Screen Function Modules

Attempts should be made to limit each function module to one screen, but the model for multi-screen function modules will effectively handle from one to four screens. For function modules that do require more than one screen, separate maps are necessary, but no other irregular changes to the model or the generated program need to be made. The primary issue is to not establish one screen for entry of key fields and another to operate on the associated data. The purpose of the banner area is to avoid this design form.

The attributes of the fields within the body of the screen are established dynamically via control variables that are set by the program dependent upon the situation. Additionally, PF Key names are set based upon the circumstances and appropriateness of each PF Key. A skeleton of a standard multi-screen function module using this technique for traditional file maintenance is:

initialize program variables
processing-loop.
   initialize loop variable
   decide for every condition
      command-id issued
         process the command
      new keys issued
         initialize new key variables
         validate action
         access primary data base record(s)
         validate data base key fields
         set prompt
         initialize NSM and screen values and screen
           validity switches
      none (no new keys or new command)
         set next prompt
   end of decide
   conditionally set dynamic field attributes and
       perform file dependencies and auto de-codes
   conditionally set pf keys
   input using map with prompt and confirmation
       based on value of screen number requested
   Process based on PFKey
   Conditional quit and return to lower level function
end processing loop
The models used by the program generator conform to this structure. See the functional specifications on the UAF Program Generator for additional information.

Mixed Case

The messages and tags on the screen should be built utilizing both upper and lower case alphabetic characters to improve readability. Error messages displayed on reinput should also be in mixed case and generally should not contain periods.

Time Stamps

All data base records should be defined with a time stamp field that is set with each record update (system variable *TIMESTMP). Update programs read the record, display the screen, and reread the record prior to performing an update. This prevents ADABAS from holding records and tying them up unnecessarily (prior to the user committing the update via PF10). The time stamp is checked after the reread and prior to the update to ensure that the record has not been updated by someone else. This approach should be used in almost all cases, but may not be appropriate if a control record or other 'count type' information is frequently updated by many keys (a class record where a count of students registered is maintained, for example).

Audit Log

Along with the time stamp field on the application, fields should also be defined that track the update history of a record. TIME-OF-ENTRY, TIME-OF-UPDATE, and UPDATED-BY-USER are three standard fields that make up an audit log. When multiple screens can update the same record, these fields should be placed in a PE group, where each occurrence is given a fixed value for each screen. That occurrence is then referenced on the appropriate screen.

De-code

Provision should be made to translate coded values which appear on a screen and display their associated descriptions. These descriptions are typically obtained by reading tables and must be provided for those users that are unfamiliar with the codes. However, this translation should only be performed upon request or when the description is already being accessed. This reduces the system overhead associated with these data base accesses to those situations where required. This function can also be used to explode an entry by displaying within a window more detail information than could fit on the base screen. PF4 is the standard means for making the de-code request.

Note: If using a window for de-code functions, it is standard practice for the only valid PF-Keys to be ENTER and PF3. No PF-Key validation rules should exist on these window maps since the window will be closed regardless of the program attention key pressed.


Alpha equivalent variables

In some instances it is desirable to allow special options for the entry of a field. A date value is a prime example since input may be desired with or without delimiters between month, day, and year and with a two digit or a four digit year. This capability is provided by allowing entry into an alphanumeric equivalent field on the screen and translating this value to the native field format desired. This extra processing and coding is justifiable to avoid use of the Natural edit mask on modifiable data and to provide multiple input formats for the same data. The edit mask is avoided because the resulting error message from data entered in an incorrect format is cryptic and does not indicate the format that should be used. Multiple input options for a value are provided as a user convenience and can greatly speed data entry when formating characters are omitted.

There are a few common considerations and standards regarding the use of alpha equivalent variables.

Definition of AEVs

Alpha equivalent variables are defined in the global data area's BKF group if the the corresponding native fields are banner keys. However, if they are not banner keys, these fields should be defined in either the MSE or a special AEV group within a function's LDA. Defining function specific alpha equivalent variables in the LDA's MSE versus AEV group is generally a matter of personnel preference; however, they cannot be in the MSE of TARGET functions.

Initialization and value synchronization

The alpha equivalent value displayed on the screen must match the value of the native field and conform to a default presentation format. This is handled in one of three ways dependent upon whether the native variable is a banner key field, an MSE element whose value only changes based upon direct user input, or an MSE element whose value may be changed by program assignments. Furthermore, due to the suspend feature, the alpha equivalent of a banner key field must be defined at the end of the SKS to ensure the proper restoration of its value.

Assignment of Banner Key Fields

Banner key fields are often programmatically set based upon user selections and options. In order for program assigned values to be displayed correctly in their alpha equivalent variables, the alpha values must be assigned before the map is presented. This assignment normally involves MOVE EDITED statements for each alpha equivalent variable. The standard for these assignments is to code them in the copycode aaCSETAE and INCLUDE this object where required. This copycode is included in different places dependent upon the model used.

Non-list models   For a non-list model, there are two possible situations where banner keys have been set programmatically and in both situations #NEW-KEYS is true. To ensure that the alpha equivalent values are correct in these circumstances, the INCLUDE aaCSETAE statement should be placed in the NKRESET modifiable block. In fact, the program generator generates this statement but leaves it commented out.

  1. Next Record processing alters the keys programmatically and sets #NEW-KEYS to true. Because only the native value of the banner key field is assigned with the MOVE HGV.#HIST-SEARCH-KEY TO SKS.#SEARCH-KEY and the MOVE BY NAME SKS TO BKF, the alpha equivalent must be assigned by the INCLUDE aaCSETAE in NKRESET.
  2. When the user comes from a list where he selected an entry that contains a banner key field with an alpha equivalent, then the selected banner key field must have its alpha equivalent field assigned. Since #NEW-KEYS is always set to true upon the initial entry into a function, and the AEV assignment must be made at this point to handle the Next Record situation described above, the INCLUDE aaCSETAE in NKRESET fulfills this requirement.

An example of a non-list function with banner key AEV fields is LVOMB in the LEAVE library.

List models   For a list model, there are also two possible situations where banner keys have been set programmatically; however, the circumstances and solutions are different.

  1. In a list model, the selection of an entry from the list results in the movement of the list key fields to the banner, but #NEW-KEYS is not set to true. If any of these list keys has an alpha equivalent, that value must be set prior to the presentation of the next screen. The program generator for list models takes care of this situation by including a generation variable that asks if an AEV field for a Banner Key is included in the LSK. If so, the INCLUDE aaCSETAE statement is generated in the section of code after map presentation that executes the move of the selected entry to the banner.
  2. The user may enter the list function from another list where he selected an entry that contains a banner key with an alpha equivalent. As with non-list models, the selected banner key field must have its alpha equivalent field assigned. Since this situation will only occur upon the initial entry into the function, the INCLUDE aaCSETAE should be placed in the INIT modifiable block. The program generator generates this statement in INIT.

For an example of a list program that contains a banner key field that has an alpha equivalent value, see TAOLDPD in the TARGET library.

Restoration of Banner Key Fields

Due to the suspend function, there is also a need to restore the alpha equivalent value for a required banner key. On the function to which the user suspended, he may change the value of the AEV banner field. To ensure that the original alpha equivalent value for that required key is restored upon the return from the suspend, that AEV field should be included at the end of the SKS fields (beyond the bytes redefined as #SEARCH-KEY) in the function's LDA. Here they will be preserved by the MOVE BY NAME BKF TO SKS so that they may later be restored upon the return from a suspend via the MOVE BY NAME SKS TO BKF.

If the above steps are followed, the alpha equivalent values displayed in the banner of all functions will always match the native values actually being used by the programs prior to map presentation.

Direct Input MSE Elements

MSE elements that use alpha equivalent variables and are only altered via direct user input should be initialized whenever the keys change. This can be done any time after the MSE values are established. It is recommended that the statements that initialize these variables (MOVE EDITED or others) be placed in the INITVAL modifiable block. From this point forward, the alpha equivalent values will be maintained in sync with the native values by the processing rules discussed below in "Entry and validation".

Program Assigned MSE Elements

MSE elements that use alpha equivalent variables that are altered via program assignments require a different solution to ensure that the alpha value displayed always corresponds with the native value. One approach is to reassign the alpha equivalent variable whenever the native value assignments are performed. However, this is sometimes accomplished in a complex routine or a protected area of code as with the PEGM model (the EdOpt changes the native values). The simplest solution is to place these assignments in the PREMAP modifiable block so they are always performed prior to the display of the map and thus always in sync with the native values.

Entry and validation

The values entered from the screen into alpha equivalent variables must be edited for the correct format and the native fields must be set to the equivalent values. This is accomplished through the use of processing rules. Invalid entries should result in a REINPUT with the text of this message customized so that the proper format for the entry is included. This processing rule should be rank 0 for all banner key fields. For banner keys, the standard is to allow the user to exit with bad data in the alpha field. The code should allow an exit if PF3, PF4, or PF5 is pressed, OR 'FIN' or 'LOG' has been entered in the #COMMAND-ID-ISSUED field but the PF-KEY is not PF10 (see UACDATEI in SYSTEM for an example). For non-banner key fields, the rank of this format validation and translation processing rule must be established with consideration for how the corresponding native field may be used in other processing rules. For example, the native value may require further validation and may be used to perform a de-code function. In this case two rules would be used, both less than rank 50, and the rule that performs the alpha equivalent format validation and translation is a lower rank than the one performing the data base validation and de-code.

Note: Since these processing rules may be less than rank 50 they will be executed during a view action if PF4, de-code, is pressed. Therefore, the corresponding REINPUT should not be performed when PF4 has been pressed.

Standard procedures

To facilitate use of this technique, to standardize input formats for common fields, and to maximize code reuse, several standard copycode objects have been created to accomplish this processing. These objects use the new Natural 2.2 feature of providing arguments to copycode so that the alpha equivalent and native variable names can vary with each invocation. These include:

UACDATEI

Allows entry of a date in month-day-year format, with or without dashes or slashes, and with a two or a four digit year. The resulting value is always displayed as: MM/DD/YYYY. The native field must be format D.

UACCCCIN

Allows entry of Company Cost Center with or without delimiting characters and with or without the project number (last four digits). The resulting value is displayed as: 9999 99999-99-9999.

UACSSNIN

Allows entry of Social Security Number with or without dashes while displaying the result as: 999-99-9999.

Please see the specific code for additional information. Also note that the last two routines accept a blank value which is translated to a null native field value (UACDATEI requires a parameter that determines if a blank value is allowed). When a non-null value is required, an explicit check for a non-null value must be provided. This can be done in the main program when dealing with banner key fields or by adding an additional processing rule for other fields.


Code translations

The maintenance of department and budgetary unit associations throughout our applications has presented a unique problem because the names and associated codes (abbreviations) used to identify these organizations commonly change. Due to this situation, numeric identifiers for these entities were created, and applications were designed to store the numeric identifier rather than the code so that massive conversions would not be required when a unit changed its name. The DEPARTMENT and BUDGETARY-UNITS tables, where possible values reside, are keyed by both the numeric identifier and the code. To aid the user in data entry the convention is to allow entry of the familiar code, translate this to the equivalent numeric identifier, and store the numeric value on the data base. Similarly, to aid the user in interpreting output, the numeric value from the data base should be translated to the alphabetic code for display purposes.

The processing considerations for code translations are the same as those for "alpha equivalent" variables previously discussed and those standards should be applied. The difference is in the manner of translation, rather than IF ... MASK and MOVE EDITED statements being used these translations require table look-ups. Two subprograms have been developed to perform this function: UANRDEPT and UANRBU. Given a numeric identifier they will return the four character code and name, or given the code (and a zero numeric identifier) they will return the number and name. Use of these routines reduces the need to define the necessary data base views and FIND statements, and therefore is encouraged. To assist in the validation of alpha codes entered online, two copy code members have been developed: UACDEPTI and UACBUCN2. These accept arguments for the appropriate field names for: the 4 character alpha code, the 4 digit numeric identifier, and the name of the unit (the name is optional; see the actual copycode for further information). (UACBUCN2 also uses a control variable to reduce I/O by only reading the file when the code field has actually been modified.) These should be invoked from a processing rule on the alphabetic code field. The same guidelines as for alpha equivalent fields regarding the rank of these rules applies to these code translations (banner key fields should use rank 0).

Code translations for key fields have some unique circumstances that are not handled the same as the alpha equivalents. The following standards have been developed to address these situations.

  1. In a list model, the user may select an entry which contains the numeric value for the alpha code in the LSK and move it to the banner. Since the code translation occurred in the database access in order to retrieve the alpha code that is displayed on screen, both the numeric value and its alpha code should be stored in the LSK. The MOVE BY NAME LSK TO BKF will then move both values to the banner so that no further database lookups are required.
  2. Code translation in a Basic Table Maintenance function (BTM model) is similar to the AEV processing for the list model: the code translation for banner keys should be done in the INIT modifiable block (see the Alpha Equivalent Variables documentation). This is because the user could have entered a code from another function and there is no Next Record processing nor TARGET processing in a BTM. The user, however, may press PF8 to page forward in the table of entries. When the user moves forward, the last table entry is moved to the banner to be used as the starting search key for the Read. If that entry includes a numeric value that uses a code translation as a key field, then the alpha code should also be moved along with the numeric value. Therefore, the alpha codes also should be stored in the Table Key Fields (TKF). The MOVE BY NAME TKF(#MAX-TBL-OCC) TO BKF ensures that the alpha code will be moved along with the numeric value to the banner and will be used by the Read.
  3. Banner key fields in a maintenance function model (MSF, TMSF, PEGM, TPEG) can be programmatically assigned by Next Record processing, transaction retrieval on a Review, or selection of transactions from a list. These situations all result in only the BKF numeric field being assigned. Since #NEW-KEYS is set to true for all these cases, the code translation lookup can be performed based on these conditions in the NKRESET modifiable block. For a non-TARGET function, the necessary code is as follows:
      IF #PS-PF-NAME = 'NextR'
        CALLNAT code-translation-subprogram
      END-IF
    
    For a TARGET function, this code should be structured as follows:
      IF (#ACTION-REQUESTED = 'R' OR TTF.TXN-SELECTED-FROM-LIST
          AND TTF.TXN-EXIST) OR #PS-PF-NAME = 'NextR'
        CALLNAT code-translation-subprogram
      END-IF
    

An example of code translation can be found in NSM-MS, command LDB (List Desks for a Budgetary Unit).


Additional Considerations


How To

The following steps describe how to develop an application using the NSM Architecture. For references to program generation, please consult the UAF Program Generator Functional Specifications for detailed instructions and a list of current models and their required components.

  1. Have the new application defined to Natural Security by the Natural Administrator.
  2. Have your application, command security groups, value security groups, application-users, and menus defined to the NSM-A files through the NSM Maintenance System (NSM-MS).
  3. Copy the model global data area (xxGLOBAL) into the application library and save it with the correct application prefix. Edit the new global to contain the required banner key fields (under the structure BKF). If the application uses value security, the structure CAF (Common Application Fields) is used to define the value restrictions. Stow the global data area.
  4. Generate the NSM-A xxOENTRY, xxOMENU, and xxOERROR programs using PGZENTRY, PGZMENU, and PGZERROR. Make modifications that are required. The entry module may be stowed at this point, but the menu program requires a map. Save the menu program.
  5. Copy the model menu map (xxMMENU) into the application library and save it with the proper application prefix. Edit it to include the banner keys defined for the application. These fields should be brought in directly from the application's global data area so that the fields contain the proper qualifier (BKF). Stow the map and then the menu program may also be stowed.
  6. If value security will be utilized, write the value security maintenance module to be included in the NSM Maintenance System. This program must be defined in the Application file as the application's value security maintenance program before users can be assigned value security groups.
  7. Have your entry module defined as the STARTUP program to Natural Security.
  8. A local data area is required for all function modules, and each function module type has its own model local data area, in addition to the one stored on SYSTEM that is required for that module type (see "Local Data Areas"). Build an LDA from the appropriate model following the method used for maps (see item 10).
  9. For each type of function module developed, a model map has been defined and stored on SYSTEM. Copy the appropriate model map (e.g. xxMMSF) into the application library, save it using the new application prefix, and then modify it to include your banner keys. At this point, it can be saved and used as a prototype map for all the same type of functions in your library, and the banner keys, off-screen elements, and PFKey processing rules will already be defined. It is desired that the banners of all maps be consistent.
  10. Use your prototype map of the desired model to create each function module map. Define the fields on the map by bringing them in from your function's local data area.
  11. Lastly, generate each function module, modify as required, and test.

Prototyping

Using the NSM Architecture and the How To steps identified above, it is very easy to develop a prototype of an application. The menus are easy to create or modify and the function module programs can be as simple or complex as desired and are easily generated. Field validations and help facilities can be added at any time. This provides a graduated means of developing an application and results in a system that a user can logon to, use and hopefully provide beneficial feedback for the development process.

Security and Auditability

Security administration of NSM-A applications is simpler and more timely than a centralized staff performing all administration in Natural Security. This is accomplished by having the application owners perform these administrative functions themselves, online, using the NSM Maintenance System. This replaces the typical procedure of:
  1. User A asks centralized security administration for access to System B.
  2. Security administration refers User A to the Owner of System B and states that written approval will be required.
  3. User A discusses his need with the Owner of System B.
  4. A memo is drafted requesting the needed access, written approval is obtained from the Owner of System B, and the request is sent to Security administration for processing.
  5. A phone call is made from Security administration to the Owner of System B to clarify or complete the information needed to establish the desired access.
  6. Security administration establishes the necessary access.
  7. A phone call is made back to the Owner of System B or User A to let them know access has been granted and that they should test out the access privileges.

Centralized security administration is entirely removed from this loop, and the owner immediately establishes the necessary access for a new user or alters that access as circumstances require. This should not be a burden since the use of the online system is simpler and more confidential than a method of sending memos and making phone calls. The NSM Maintenance System also provides a record of who last made a change (See "Audit Log"), allowing some means of investigating any disputed entries.


Conversion to NSM version 2

There are several steps involved in converting applications to the second version of the NSM architecture. Each of the following conversion instructions should be followed in order. For all references to program generation, consult the UAF Program Generator Functional Specifications for more information on that facility.

The basics: GDA, Entry and Menu Programs

Create a new Global Data Area as described in the section "How To", remembering to include your key fields and value restrictions (if used). Generate the Entry and Menu programs.

Individual Function Modules

It is recommended that the programmer convert function modules in groups of model types. The logical first step in this process, therefore, is to determine into what model category the programs fall. This procedure necessitates that the programmer be familiar with the developed models.

Following the instructions in "How To" and the program generator documentation, construct your function's template local data area for the specific model (e.g. xxLMSF).

Function Module Maps

The maps are constructed following the rules in the UAF Program Generator Functional Specifications. For the conversion process, the most efficient method to convert maps is usually to use the old map as the foundation for the new one. In this way, the inline field validation rules will not have to be re-typed. The following list details the steps for conversion of maps.
  1. Delete the old Title field and replace with new TITLE-AND-CONFIRMATION array field from the local data area (stored on SYSTEM) for the particular model with which you're dealing. It is important that this field be set up properly so that the title and confirmation will be displayed correctly. Define the array (".a" command) as one occurrence beginning with #TAC-OCC. Also, put in the Dynamic Attribute parameter (DY) the following: ^I|, which will allow the confirmation to be displayed intensified and the command-id at the end of the field to be displayed default intensity.
  2. Delete remaining banner fields and bring in new ones from the global data area, using the BKF structure fields. Remember to put in the proper help routines. Control variables are now used with the banner key fields (and all these fields are Output Modifiable -- COMMAND-ID-ISSUED used to be Input but is no longer): #MI-FA is used as the control variable for those key fields in the banner that are to be intensified, and #MD-FA is to be used for those fields to be default intensity.
  3. Fields directly below the banner that show the user the action and record being acted upon also need to be changed. The Action field that was ACTION-IN-PROCESS is now CFF.#EFFECTIVE-ACTION (included from the model's lda), and the key fields being acted upon are now the Search Key Structure (SKS) fields (included from your function's lda). See any of the NSM*** maps in NSM-MS for examples.
  4. Delete the off-screen elements which were standard to NSM version 1. Replace them with the standard fields for the particular model with which you're working.
  5. Unlink the old PFKey processing rules, and include the new rules for the specific model. Unlinking is accomplished by typing 'unlink' on the command line of the processing rule editor. All standard verification rules begin with the prefix 'NS-'.
  6. Rules that perform decode functions should be of rank less than 50. Likewise, validation rules that do not need to be performed if the user presses PF4 should be of rank 51 or greater.
  7. Change the qualifier (normally READ-VIEW.) to the appropriate qualifier. For example, for MSF functions, the qualifier is MSE. This can most easily be done by typing 'D' (Display Field Definitions) on the main map editor menu and going through the list and replacing READ-VIEW with the correct qualifier.
  8. Decode fields normally are qualified with DCF.
  9. Rename map names for multi-screen functions to conform to the standard format (See "Map Name Format").
  10. Ensure that the map contains all the new fields (like CFF.#SCREEN-REQUESTED for MSF) required for the particular model and that the necessary control variables are properly used.

Generate, Modify, and Test the Program

After the function module's local data area and map(s) have been created and stowed, generate the program using the appropriate edit macro. Modify the areas that are required, and change or add any other code needed. Finally, thoroughly test your program.