04 September 2022
اخبار انگلیسی
Although most changes can be costly and unreliable and ultimately lead to customer loss, modern businesses in today's world want to change systems to provide better products and services. The problem with modernizing mainframes is that the tools used to search code and analyze programs have drawbacks in reducing the risks associated with maintaining and improving systems.
Modernization and Knowledge Management
Given that there are currently 10,000 active mainframes worldwide, an organization should spend an average of 60 to 80 percent of its IT budget annually on maintenance. But the main problem is that with the retirement or departure of experienced and skilled programmers and developers, their specialized knowledge also leaves the organizations and companies to create and update mainframes, which are complex and vital systems, and will face problems. Therefore, programmers should document all their work from the moment they enter the organization so that the organization can use their knowledge.
Developers who have been working in the organization for a long time leave a lot to be documented. This allows new developers to gain a comprehensive knowledge of system performance.
Today, developers spend about 75% of their time searching for source code. To maintain and support any system effectively and efficiently, software developers need to know what an application is working and how changing the principle of one part will affect the whole system. But because similar codes might be all over the system, developers may think they are making a simple change when it may crash the entire system. So, they need to be aware of it.
Tools and Problems
The number of code repositories is increasing day by day. Since there is no monitoring and control over them, those responsible for maintaining the system's performance believe that finding errors in the code without the help of AI and ML will be challenging. As debugging is a laborious and time-consuming endeavor, developers increasingly turn to static and dynamic code search tools, liners, and analytics tools to analyze millions of lines of code to mark the errors and have solutions for them.
Even if the tools are the best for troubleshooting, code virtualization, or error detection, programmers ultimately need to use their knowledge to solve the problem. Application analysis tools display the code in various forms, and developers still have to conclude. Although code search tools may speed up the creation of a mental model of regulation, they have their drawbacks, and developers should identify the code and then make the changes. Tools used as code complements by inexperienced developers may suggest incorrect changes (a prominent tool provides incorrect code 71% of the time), which can be problematic.
Today's tools are not very efficient, and people have to change according to their goals. Ideally, solutions with fewer risks should be used to alter older systems. In this regard, artificial intelligence can be practical. Using a unique approach with the help of AI technology, the code repository becomes a repository of knowledge, enabling many capabilities that were not previously realized.
AI-Based Repository
Using AI technology to automate code line recognition, developers can easily use AI-based knowledge repositories to identify code quickly. This puts artificial intelligence in a position to become a vital need for any organization.
The new approach to artificial intelligence technology re-analyzes what the computer shows and turns it into understandable concepts. So it thinks the same as humans who think about the code. The result is that software developers no longer need to discover the purpose of previous developers and can use artificial intelligence to identify problems and fix them before compiling or reviewing the code and simulate changes and display results after changes.
Of course, we should keep in mind that artificial intelligence is not yet advanced enough to make safety changes on its own. Still, as a repository of knowledge with different capabilities, it can guide developers to get exactly where they need to go. Slowly but surely, with AI, programmers can confidently make any necessary corrections without damaging the entire system, even on the first working day.
Unlike the tools used for code search, code navigation, code completion, and program analysis, AI essentially allows developers to make any necessary, efficient, and practical changes to have more time for other stuff.
