My latest story is about the mathematician June Huh, who came into the field late and by an unorthodox path. Huh’s approach to mathematics has been similarly surprising: He and his two collaborators, Eric Katz and Karim Adiprasito, solved an important problem called the Rota conjecture by figuring out how to translate ideas from one area of math to a realm where those ideas wouldn’t seem to belong.

Breakthroughs in mathematics often come by surprising means — problems that seemed hopeless become solvable when mathematicians find a new interpretation for established ideas. This means more than finding a new use for an old tool. It’s really about extending a pattern of argument that arose in one setting into another.

[abstractions]

A good example of how this kind of translation works involves a concept called the Laplacian. The original Laplacian was developed by the 18th-century scholar Pierre-Simon Laplace to understand planetary motion and is constructed out of the second derivatives of a function. That function, for example, might describe Jupiter’s trajectory in space, in which case the Laplacian of the function would tell you something about its gravitational energy at a specific time and place.

The Laplacian arose in the world of continuous mathematics, where functions describe smooth, unbroken spaces. In discrete mathematics, by contrast, you have objects like graphs, which are collections of points (vertices) glued together by sticks (edges). Yet there is such a thing as the “graph Laplacian,” which follows from certain non-obvious properties of the original Laplacian. It’s not really a derivative, and it’s not calculated in the same way as Laplace’s Laplacian, but it retains its spirit.

To calculate the graph Laplacian, first imagine a graph and also a function on that graph that outputs a value at each point. For example, maybe the points of the graph represent discrete points in Jupiter’s orbit, which can be input into a function that tells you Jupiter’s velocity at those points. The Laplacian of that function is a new function, which assigns new values to each point (just like the derivative or second derivative of a function assigns new values to each input).

[text]

Here’s how you calculate it. Say point *A* is connected to points *B* and *C*. To calculate the value of the Laplacian at *A*, first add the values of the function at neighboring points, then subtract the value at *A* times the number of ways *A* is connected to other points.

On a graph, subtracting the values of connected points is the appropriate analog of taking a derivative, which is, after all, the difference of two values infinitely close together. And given that, differentiating a function becomes like taking the difference between two points connected by an edge. (For a particular reason related to something called Green’s second identity, this subtraction operation mirrors the second derivative, rather than the first.)

The correct analogue of an established mathematical idea is often not obvious, and finding it can take awhile. But once established, it becomes quite powerful. Huh, Katz and Adiprasito spent four years searching for the right way to translate certain ideas into the realm of graphs. Once they’d found it, a proof of the Rota conjecture followed in a matter of a few months.

Projects produce products, services, solutions. Projects deliver capabilities that fulfill Missions and Vision, Projects are a vehicle to accomplish and outcome. Projects collect *nonrecurring* costs in exchange for that produced value. Every project should have a goal, an objective, and outcome. Projects in various domains have different practices and processes. But each of those should be anchored on a set of Immutable Principles.

Here's a set of Principles, Processes, and Practices that have served us well for a decade or so.

So let's look at these in more details

The Five Principles of Project Success

**What Does Done Lool Like?**It's popular in the agile work to claim*SW is never Done*. That may be a nice idea, but from the balance sheet Point of View, Done means we can record the capital expense or we can record the Capital Income from the software. FASB 86 speaks to this as do other accounting practices. Now we may have continuing versions of the software, but at some point*done*means we are shipping a version to those who paid for that version (even if they are*renting*the software) and we have recorded that as an event. In the Change Control Paradigm, that version has a version number, it has a branch in the version tree, and we are not changing that code base without updating the version number. That version is Done. Projects produce outcomes.*Done*needs measures. These start with*Measures of Effectiveness*and*Measures of Performance.*These need to be written down before starting, otherwise, you won't recognize then when or if they appear.**How Do We Get to Done?**No matter the approach to delivering outcomes, some sort of Plan is needed. It can be a formal Integrated Master Plan, a Product Roadmap, and checklist, or something in your head. But a Plan is needed. Plans are strategies. Strategies are Hypotheses. Hypotheses need to be tested to confirm they will hold up in practice. So Plans are meant to be tested and Changed when found to be wanting. Anyone suggesting plans and planning is waste has not be held accountable for any non-trivial deliverable. The Plan tells us what to expect at specific points in the project.*We Plan to have the first instance of the customer enrollment systems working by the end of the First Quarter, so we can switch over to the new customer tracking system*. Or*we Plan to have the heat shield integrated with the spacecraft ready for the static testing down at the Cape by the end of the summer*. A Plan states*intent*and a date when that*intent*will be true.**Do we have everything we need for the Plan to produce Done?**- resources include everything need by the project. People, money, facilities, technology. Anything needed for the success of the project. These resources need a plan. A time phrased plan for their acquisition, deployment, management, and funding.**What could possibly go Wrong?**- risk management is how adults manage projects - Tim Lister. This is the basis of managing projects. All risk comes from uncertainty. Uncertainty comes in two forms - epistemic and aleatory. Epistemic uncertainty is reducible. Work can be applied to reduce the uncertainty. Aleatory uncertainty is irreducible. Only margin can be applied to this uncertainty. Schedule margin, cost margin, technical margin.**How are we going to measure progress to our Plan other than the passage of time and consumption of Money?**- Physcal Percent Complete is the only means of measuring progress to plan. The defined physical percent complete at a pre-defined time and pre-defined value. When that time comes, the measure of physical percent complete is assessed and any variances used to make corrective actions.

So in the - or actually the beginning - the Five Immutable Principles all need to be in place before the probability of project success can be increased

SAFe 4.5 can be configured for four development environments (Essential SAFe (simplest), Portfolio SAFe, Large solution SAFe and Full SAFe (most advanced) and SAFe 4.5 is backwards compatible with SAFe 4.0 (available through June, 2018).

The big picture is a little bit more user friendly (some details / icons are taken out) and you can select the environment you want to use (and as a result the picture will only show the corresponding parts). I think some more icons on the big picture can be removed too if Scaled Agile creates a specific big picture on their homepage for their online knowledge wiki which contains hyperlinks to all topics. A big picture to explain SAFe can work without icons for SPC, Lean-Agile Leaders, the implementation Roadmap and topics like the Continuous Delivery Pipeline.

Most important changes:

- Faster innovation with Lean Startup and Lean UX
- Epic Value Statement and Light weighted Business Case are replaced by Epic Hypothesis Statement and Lean Business Case
- Feature delivery with Scalable DevOps and Continuous Delivery Pipeline
- SAFe implementation Roadmap
- Value Stream has been changed to Solution (Value Stream Backlog > Solution Backlog, Value Stream Engineer > Solution Train Engineer, …) and a Solution Train has been added (several ARTs and Supplier forms one Solution Train)
- Compliance has been added to the Solution Intent
- Program Portfolio Management (PPM) has been replaced by Lean Portfolio Management
- Increased alignment with the Scrum Guide

Impact on my book *Scaling Agile in organisaties* is minor. For the coming year, SAFe 4.0 is still valid. In the next print of my book I can make adjustments regarding the name changes (Value Stream, PPM). The SAFe Implementation Roadmap was already incorporated and topics like faster innovation with Lean Startup can be added.

More information on SAFe 4.5 can be found on www.scaledagileframework.com. At the homepage you can with to SAFe 4.0 too.

Next Page of Stories