Best Practices for Appian developers
Appian best practices are like a check list, use these checklists at different stages of your development cycle as described to ensure application success and reduce technical risk. Each section provides guidance for when and how often you should use these.
5.Best practices for Creating Process Model :-
1.All process models should follow the naming convention using the form “Verb Noun” prefixed by application name.(Example <APPPrefix> Create User)
2.All process models should have security defined at the lowest security level as possible.
3.All process models should have dynamic display names.
4.All process models should define shortest archiving period as much as possible.
5.All process models should have alert configured to the application administrator or process designer groups.DO NOT configure alert to individual users.
6.Process models should have proper Swim lanes and Annotations.
7.Swim lanes task assignments to be used only if all the tasks are to be actioned by a single group
8.Process models should not have more than 30 process nodes.
9.Process nodes should follow the “Verb Noun” naming convention (Example: Get Task List)
10.Unattended process nodes should be assigned to execute in context of designer instead of initiator
11.Process models should not have more than 50 process variables.
12.Process Variables should follow camel case naming convention.(Example: listRecords).
13.All the process models needs to create in a specific folder related to the application.
14.Provide Folder and process model security to access the process models
15.Provide display name and the description for every node present in the PM.
16.Set the priority as normal for all nodes
17.Process nodes should be visible within the canvas of process model designer and the node connectors should not overlap.
18.Should label all gateway nodes in the process model in the form of questions.
19.Label all the outgoing flows from gateway.
20.XOR/OR gateways should have single incoming flow.
21.XOR gateways are preferred over OR gateways as much as possible.
22.Two gateways should not connect together.A script task or a dummy node can be used in between two gateways
23.Scripts task to be configured only with Data Output and not Data Inputs
24.Configure all the task display names in the UI by means of expression rules.
25.Configure Assignments in the UI node by assigning groups or a process variable containing the assignee.DO NOT configure task assignment to a user directly.
26.Avoid using Deprecated functions.
27.Use Query Entities / Write to Data store nodes instead of Query Rules and Query Database node.
28.No CDTs should be configured as "Pass by reference" in the inputs of the sub process.
29.Save Utility process models in the separate folder.
30.Exception path has to be set using expression rules and constants for User input Task.
31.Timer values should be stored in Constants and use the constants for all timer configurations in the process.
32.Use SAIL wizard instead of activity chaining with multiple user input tasks.
33.Enable “Delete the previously completed/Cancelled instances” for nodes that will be executed in loops
34.Designers should not reuse process models across applications.
35.If a PV is passed into a sub process, the sub process and the parent process should have the same PV name.If they do not have the same PV name, task reports will not be able to show the data correctly because tasks in the sub process cannot see PV names in the parent process.