In the current Internet era, more and more companies are beginning to use network-based, information-based integrated workflows to help themselves improve operational efficiency.
In order to solve the problems in various stages of the enterprise, the workflow naturally develops and develops with the times. Nowadays, mobile Internet has become popular, and the computing power of each terminal device is getting higher and higher. Many companies have begun to require the management of workflow to be mobile and intelligent, and compatibility must be guaranteed to be able to handle business anytime and anywhere.
1. Trends in workflow modeling
The current modeling technology of workflow can be described as a hundred flowers blooming, there is no unified model, which means that in terms of technology, related solutions are still not mature enough. In summary, the current mainstream modeling methods can be basically divided into five categories: scripting languages, web-based methods, rule-based methods, logic-based methods, and algebraic methods.
(1) Script language
Scripting language naturally goes without saying, it usually includes the statement structure of control flow and data flow. The advantage is that it provides a relatively simple description method, which is more handy for experienced designers. Of course, the shortcomings are also obvious, that is, it lacks the formal semantic meaning of the process itself, and the semantics of the language is mainly used by the language interpreter.
(2) Web-based approach
It is suitable for process visualization modeling, generally using state transition network, the activities in the process are represented by nodes, and the control flow is represented by edges.
When using the state transition network, you need to pay attention to whether it has formal semantics. Most of the visual modeling methods of workflow products lack formal semantics. Among the state transition nets with formal semantics, the Petr nets and State Charts are the most used. Web-based methods with formal semantics can be converted into other modeling methods, such as rule-based, sequential logic, and scripting language methods.
(3) Rule-based approach
At present, the more mainstream rule-based method is ECA (Event/Condltion/Acton) rules. ECA rules were first used in AooDBS, and then used in the field of workflow management. ECA rules have a formal basis and can also be converted into other modeling methods. However, the visualization workload of ECA rules is relatively large, and it is difficult to manage when the rule set is large.
(4) Logic-based approach
Logic-based methods are suitable for describing the dynamics of the system. Temporal logic is a commonly used method. It has a good formal foundation and is more convenient to verify the attributes of the workflow model. But the main disadvantage of temporal logic is that it is difficult to realize visualization. It is not easy to convert to other description methods, and the system behavior of describing business processes is too complicated.
(5) Method based on algebra
Process algebra (Proees Algebra) is mainly limited to theoretical discussion, and is rarely used in the field of workflow management. Only a description language LoToS based on process algebra is used in the field of workflow management. The main disadvantages of the algebraic method are similar to the logic-based method. Its performance in automatic execution and formal verification is worse than that of the logic-based method. The modeling method lacks intuition and is difficult to understand.
According to the above comparison, it can be seen that various modeling methods have their own advantages and disadvantages. But in general, scripting languages, web-based methods, and rule-based methods are more attractive.
2. Environmental trends for workflow implementation
Workflow management systems should support the integration and interoperability of application systems in heterogeneous, autonomous and distributed environments. Provide methods to integrate old application systems to protect past investments and support the reorganization of organizations. At the same time, it is also necessary to support dynamic enterprise technology with fault tolerance, and the workflow management system can guarantee work when errors occur The correctness and reliability of flow execution.
Later, some web-based workflow solutions appeared. By observing the workflow management system based on Web technology, it is found that most products can only use the Web one-sidedly, but facing the Web is the future development trend. This trend can be reflected in some research projects (WebFlo, OzWeb, DartFlow).
Due to the limitations of the Web and the browser itself, only the Client/Server computing mode can be provided, and the CGI interface used has only limited programming capabilities. Further improvements are needed in terms of location transparency, transaction support, security, and performance. . And workflow research is a cross-disciplinary research involving CSCW, human-computer interaction, database, management, sociology and other disciplines. Any research that lacks multidisciplinary cooperation cannot make it a universal system.
In the choice of workflow, companies usually need to commission software companies or self-organize technical teams based on their own business characteristics. Generally, only functional requirements are mentioned when commissioning, and there are no too many requirements for the realization mode. The worry is that it saves worry and effort, but the disadvantage is that it is difficult to maintain in the later stage.
There used to be a company that purchased a finished product BPM and needed to upgrade the system after business changes. However, the company lacked relevant technical personnel and could only find a development company to spend money to upgrade again. The upgrade was a lot of trouble. Even if you have your own developers, upgrading existing software is a big project. After all, no one wants to modify other people's code, and there are too many problems involved.
When there are problems, some people naturally think of ways to solve them, so the rapid development platform came into being. It is not a traditional finished software, but a semi-finished framework that integrates various functional components, which can be developed twice. Enterprises only need a small number of developers, and do not have too many requirements on the developer's technology.
For example, the relevant frameworks I have come into contact with basically include code generators, workflows, forms, permissions, reports, WeChat integration, etc. Simple configuration based on the frameworks can quickly develop including OA/ERP/CRM/mobile APP/ The enterprise application system including the e-commerce backend is basically source code delivery, and subsequent business changes can also be handled by itself, and many things will not be controlled by others.
Here is a demonstration demo of a partner is recommended. You can first understand the basic routines of rapid development from the basic architecture, and then select based on the company's own situation.
Demo address: https://www.xjrsoft.com/