THE LINUX FOUNDATION PROJECTS
Blog | Mainframe Open Education | Mentorship

Summer Mentorships 2026: Halfway Through My LFX Mentorship – A Platform and a Blog Series for Mainframe Open Education

By | July 29, 2026

 

Tushar BiswasWritten by Tushar Biswas, Open Mainframe Project Summer Mentorship 2026, mentee researcher, guided by mentors Kelsey Colberg, Nanxi Meng, and John Lovett

I’m six weeks into a twelve-week LFX Mentorship with the Open Mainframe Project, working on the Mainframe Open Education (MOE) initiative alongside my co-mentee Idris, under the guidance of Kelsey Colberg, Nanxi Meng, JJ Lovett. This is the midpoint check-in — what got built, what got written, what I got wrong along the way, and what’s still left to do.

I want to be upfront about the format: this isn’t a highlight reel. It’s a working log, including the parts that aren’t finished yet.

The problem we’re actually solving

Mainframe knowledge isn’t scarce. Decades of documentation, IBM Redbooks, vendor manuals, and institutional knowledge exist and are genuinely excellent. What’s scarce is a good on-ramp — something a student or early-career engineer can sit down with, actually run code in, get quizzed on, and walk away with proof they learned something. Most mainframe content today is built to be referenced by people who already know what they’re looking for, not discovered by people who don’t.

That gap is what MOE is trying to close, and it’s what most of my work this summer has gone toward.

What shipped: The MOE Learning Platform

The bulk of my time went into building an interactive platform on top of the existing MOE GitBook content, rather than treating the content as a static reference. It’s live now, and includes:

  • A live COBOL sandbox using the Piston public execution API, so readers can run real COBOL in the browser instead of trusting a static code block.
  • A 130-question quiz system, generated per chapter, aimed at testing comprehension rather than recall.
  • PDF certificate generation via Puppeteer, so completing a module produces something tangible a learner can keep.
  • A scroll-driven journey map, styled around a JCL-console aesthetic — a small design choice, but one that signals the platform takes the subject matter seriously instead of wrapping it in a generic ed-tech skin.
  • Cmd+K search, dark mode, and progress tracking — unglamorous, but expected by anyone using a learning tool in 2026.

The stack is intentionally boring where it should be: Node.js/Express, Docker, deployed on Render. Boring is a feature when the goal is that other contributors can pick it up without a ramp-up period of their own.

What I wrote: the Mainframe 101 series

In parallel with the platform, Idris and I have been co-authoring the underlying curriculum. My chapters:

The hardware piece runs close to 2,800 words covering IBM, Unisys, HPE, Fujitsu, Hitachi, and NEC. The operating systems piece is a 4,000-word comparison of five genuinely different design philosophies running on the same class of hardware. Writing the z/TPF section in particular changed how I think about the platform — millisecond-level transaction guarantees for global airline booking aren’t legacy inertia, they’re a deliberate engineering answer to a scale problem most modern stacks never encounter.

Publishing runs through Medium first, with Dev.to and LinkedIn as distribution channels — the goal being to reach engineers who’ll never open a GitBook on their own but might land on the content through a normal feed.

What’s still open

In the interest of transparency over polish, here’s what hasn’t shipped yet:

  • Admin panel and analytics dashboard. I know the platform works. I don’t yet have visibility into how people are actually using it, which is the next real gap.
  • Credly integration, so completing a module produces a portable, verifiable credential rather than only a downloadable PDF.
  • Email-based login and certificate storage, so progress persists across sessions instead of resetting.

None of these are small, and I’d rather say that plainly than round up on progress.

What I’d tell someone starting something similar

The biggest shift in how I think about this project happened around week three: I stopped treating this as a content project and started treating it as a distribution problem. The knowledge already exists and has been battle-tested in production for decades. What doesn’t exist, in most cases, is a version of it that meets a new learner where they actually are — in a browser, wanting to run something real within the first five minutes, not the fiftieth page.

If you’re contributing to open source documentation or education projects, that’s probably the more useful question to ask early: not “is this accurate,” but “will someone with zero context actually get through it.”

Thanks, and what’s next

Six weeks left. The priorities for the back half are the admin dashboard, Credly integration, and closing out the remaining curriculum chapters. Thanks to Kelsey, Nanxi, JJ, Lauren and Linux Foundation for consistently pushing on both the technical bar and the “who is this actually for” question whenever I got too deep into implementation details and lost sight of it.

More at the end of the term.

 

Stay tuned for more mentee blogs as our Summer Mentorship Program continues!

Share