IBM Just Killed Merlin. Here’s Why

October 20, 2025 Alex Woodie

The writing had been on the wall for some time, and IBM finally made it official as part of the Technology Refresh for IBM i on October 7. That was when Big Blue announced the immediate withdrawal from marketing and service for its Modernization Engine for Lifecycle Integration product, commonly called Merlin. It will likely go down in IBM i history as a well-intentioned product that got some things right, but which ultimately could not overcome its flaws.

The introduction of Merlin back in May 2022 represented a sizable shift in IBM’s approach to application development on IBM i. Prior to that, Rational Developer for IBM i (RDi) was the pretty much the only modern way of developing traditional ILE applications for IBM i. The Java-based integrated development environment (IDE) was well-established and feature-rich, but it was also pricey and slow. A large chunk of the IBM i developer population used RDi, but a sizable contingent also preferred to stick with their greenscreen tools, like SEU and PDM, thankyouverymuch.

The future of application development, IBM brass correctly predicted at the time, lay in lightweight and open Web-based IDEs. By the late 2010s, Microsoft’s Visual Studio Code had already established itself as the standout IDE in this category. As a free and (mostly) open source version of its full Visual Studio IDE, VS Code was widely adopted by the community, with thousands of supported libraries and plug-ins for a variety of development tools, languages, and runtimes. As a relatively new IDE, VS Code natively supported DevOps methodologies and technologies like Git and Jenkins, as well as cloud-native deployment of applications via containers and microservices. It already had many millions of users by 2019.

IBM i chief technology officer Steve Will saw everything going on with VS Code and wanted it for IBM i too. He saw that a large part of the IBM i user base was still immersed in legacy development tools and methodologies. They were using SEU to maintain older green screen apps developed in RPG III and legacy RPG IV and DDS code, rather than adopting newer languages like free form RPG IV and SQL. Will reckoned if he could give them an open, lightweight, Web-based environment for developing modern free form RPG applications, one that was fully integrated with DevOps technologies and supported containers and microservices, that they would respond positively.

These were the general goals that IBM trumpeted when it formally unveiled Merlin to the world at the POWERUp 2022 conference in New Orleans, Louisiana. However, Merlin wasn’t just one discrete thing. Instead, it was a collection of disparate tools and tech, which ultimately was one reason it was cast out of the kingdom.

For starters, instead of VS Code itself, IBM decided that the core elements of the IDE would be Che and Theia, two Eclipse-based open source projects. Che replicated the VS Code development space, while Theia was a tool for developing IDEs. IBM selected these components, as IBM i business architect Tim Rowe explained to us last year with the launch of Merlin 2.0, because VS Code was not fully open source at the time. IBM would drop Che with Merlin 2.0 and adopt VS Code outright after Microsoft open sourced the VS Code development space, Rowe told us, but by then it was probably too late.

On top of the VS Code-like core, IBM brought in some other tools. It leaned on its partner ARCAD Software to provide integration with Git for source code management and Jenkins for continuous integration/continuous deployment (CI/CD) pipelines. It also adopted an RPG conversion tool from ARCAD called Transformer, which IBM renamed Converter for its Merlin offering. The partnership with ARCAD was exclusive, which irked some of IBM’s change management business partners who were not invited to Merlin’s round table.

IBM put one more trick into Merlin’s bag of spells: Red Hat OpenShift. Since Kubernetes was (of course) the future, IBM decided to require Merlin to be deployed as a container within the OpenShift Kubernetes platform. This meant Merlin needed a separate Linux environment, which could be running in the cloud, in a separate Linux box, or in a Linux partition running on a Power Systems server. But it wasn’t running natively on IBM i, and we all know how well that works with a certain contingent of the IBM i installed base, particularly the portion that has so far resisted modernization (Hint: it doesn’t work well).

IBM decided to bring OpenShift into the Merlin equation for a couple of reasons, according to Alison Butterill, who was the IBM i product manager at the time. For starters, it allowed IBM “to encapsulate a lot of this function that we wanted to deliver and to manage it in a very streamlined fashion,” she told us back in 2022. Secondly, it allowed IBM to “feature some of the OpenShift function,” she said. (In other words, IBM decided to use Merlin as a vehicle to force IBM i shops to adopt Kubernetes, which is the future, of course.)

To quickly recap, IBM did a lot of stuff right with Merlin. The company correctly ascertained that open, lightweight, browser-based IDEs would be the preferred environment for developers going forward. It also correctly determined that moving to DevOps tools and techniques is non-negotiable for a modern business platform. These are elements that could have made Merlin a success.

However, IBM made a few missteps with Merlin, which ultimately were the source of its downfall. It deployed a mix of open source and proprietary offerings and charged a sizable amount of money for them. It didn’t choose VS Code out of the gate, but then tried to pivot to VS Code at the last second. It excluded change management business partners. But worst of all, it forced IBM i shops to adopt a technology – Kubernetes – that is not supported native on IBM i (and cannot be) and which has no strategic value to IBM i shops.

The good news is that all of the modern development capabilities that IBM tried so hard to deliver with Merlin are available by other means (except for the Kubernetes part). VS Code has become one of the most popular IDEs in the world, and IBM is fully supporting VS Code through the open source Code for IBM i (Code4i) plug-in.

Liam Allan created Code4i back in 2021 while working as a consultant at Seiden Group. By November 2022 – just six months after Merlin 1.0 was launched – there were already 10,000 installs of Code for i from the GitHub repo where the project is managed. Today, the project boasts more than 60,000 installs, per its page on the Visual Studio Marketplace.

Merlin never came close to matching those numbers. In fact, Code4i has all but certainly surpassed RDi as the most popular IDE among IBM i developers. It sat one percentage point behind RDi – 54 percent versus 53 percent –according to the 2025 Fortra IBM i Marketplace Study survey, which was conducted in late 2024. Merlin, by contrast, had a 2 percent share, up from 1 percent the prior year. If Code4i does not surpass RDi when Fortra publishes the results of the 2026 IBM i Marketplace Study in a few months, it will come as a surprise.

IBM hired Allan just a few months after it launched Merlin back in the spring of 2022. Allan has been instrumental in helping to marshal IBM engineering resources towards Code4i. Since Allan came on board, IBM has committed real dollars to building out the Code4i toolset, including through the Code4i database extension and the Code4i testing extension that IBM is shipping with IBM i 7.6 TR1 and 7.5 TR7.

Clearly, IBM sees Code4i as the best vehicle for delivering these capabilities. Merlin may have benefited from some of the Code4i development work, such as the database extensions and the debugger, but why maintain duplicate vehicles when one of them is wildly successful and the other is not? IBM may not be making a lot of money on Code4i – although it did just launch a professional tech support plan for the product – but it is delivering the modern application development functionality that the IBM i market actually wants. That is not only enough, but it’s what IBM should do.

Will acknowledged the role that Code4i’s runaway success had in the cancellation of Merlin, while also bringing up the latest AI-powered IDE, codenamed Project Bob, which will function as a VS Code plug-in, and which Allan is now working on.

“The primary reason [for sunsetting Merlin] was the more streamlined focus on where the market is going,” the IBM i chief architect and Distinguished Engineer told IT Jungle last week. “So we know that there’s been standardization around IDEs like VS Code and where we’re going with the code assistants. And so that that was really the main reason. We’re just sharpening our focus on those things, where the market is taking things.”

You can read IBM’s Merlin announcement here.

This Issue Sponsored By

Table of Contents

Content archive

Recent Posts

Subscribe

Pages

Search