“Show me your faith without deeds, and I will show you my faith by my deeds.”
— Book of James 2:18, NIV

‍

‍

Remember that moment in Indiana Jones and the Last Crusade when he encountered a bridgeless chasm deep in the bowels of Petra? On the other side was the Holy Grail, which would save his mortally wounded father. Does he jump, or does he let his father die?

‍

Modernizing mainframes can feel like an equally poor set of choices: do we limp along with our patches and duct tape, or do we select an IT services vendor on the basis of some hand-wavy case studies and a free Paddock Club ticket to an F1 race?

‍

Last month, we wrote about the problem with traditional modernization RFP responses, part of which is that most vendors have no way of offering anything close to a guarantee of final outcomes. They know this, too: despite a surge in jargon around “outcomes-based pricing” a decade ago, the real challenge of defining a truly valuable outcome (and the asymmetric burden placed on vendors) means that even now, 90% of IT services companies don’t include outcomes-based pricing in their mix of models, according to Gartner’s Tech CEO Benchmarks Survey*.

‍

Add in a dearth of any other approach, critical buying decisions around modernization often come down to intuition, brand perception, and personal relationships. Another homo economicus sacrificed on the altar of bounded rationality.

‍

So, what really would be the “measurable outcome” when it comes to modernization? Surveys largely boil it down to these:

1. Remove constraints (often called “business agility” or “enable future growth”)
2. Find cost savings in maintenance and operations
3. Migrating without disruption

‍

Whoa, big wish list there. So let’s break it down: what would be the atomic unit that would indicate definitively that you’re on your undisrupted, controlled, provable way to a set of modernized applications that are free of historical constraints (and cost structures)?

‍

We posit that it’s working software; specifically, an incremental slice—let’s call it a workload—that is written in modern, idiomatic code and proven to be equivalent in function and performance with what it’s replacing.

‍

Do this, and you can even start thinking of win-win pricing that takes the form of a “success fee” for every increment of the application that’s working. Success is objective: it either works or it doesn’t. That, in turn, yields certainty, control, flexibility, agility within the modernization process.

‍

We all can’t be as lucky as Indy, who decoded the ancients’ metaphor and stepped onto a bridge hidden by forced perspective.  But we can take a series of straightforward, successful steps toward the “holy grail” of reducing the cost of change, in perpetuity. Because it’s a heck of a lot easier to have faith when you can see things working in non-mysterious ways.

‍

News and Views

What’s the difference between a “bag of heuristics” and causal reasoning capabilities? Christopher Mims reports on the initial conclusions of interpretability—and AI’s fallibility in handling exceptions to the rule. The Economist further digs (subscription required) into interpretability techniques as a way to avoid unintended consequences.

‍

Hard to resist a well-documented rant, and Kaushik Srinivasan has delivered a zinger. In “The Emperor Has No Clothes: The Inconvenient Truth About GCC and Leadership,” he writes an amply footnoted piece about the rise of performative IT services.

‍

Jordan Anderson explores in his article, “Data Engineers Should Be Held To The Same Standards as Bakers,” the corrosiveness of low expectations and holds forth on why developing software products on time, under budget, and happily accepted isn’t in fact the norm.

‍

Samuel Arbesman has written a book that delights “in the wonders and weirdness of computation” called The Magic of Code. His thesis is that software is inherently multi-disciplinary, and “is capable of revealing connections to biology, human language, physics, society, and even philosophy.” Yes, please. Preorder here.

‍

Fascinating interactive picture of global migration derived from 3 billion anonymized active Facebook users.

‍

From the Orchard

We debuted our modernization platform, Imogen, earlier this month. A few of our favorite coverage pieces include Network World’s “Mechanical Orchard taps gen AI for mainframe-to-cloud modernization,” Leda Glyptis’s Fintech Futures column, “What would you do if there was no gravity?”, and the blog from the Google Cloud team, “Accelerate Mainframe Modernization with gen AI from Google Cloud and its partners.”

‍

We’ll be at the Gartner Applications Innovation and Business Solutions Summit in Las Vegas. Our CCO Edward Hieatt will be speaking on June 4th at 12:50pm PDT in Theater 2, about changing the risk equation around legacy modernization.

‍

Subscribers may have already seen our special edition introducing Imogen and announcing our first partnership with Thoughtworks. We reflected on how far we’ve come, CEO Rob Mee shared how we got here (and what’s next), and Edward followed up with why business outcomes, not activity, must anchor real change.

‍

Curious to learn more? Say hello@mechanical-orchard.com.

‍

——————

Issue first published on April 29th, 2025.

‍