A few(?) more thoughts on the battle between Scrum and Software Development. A review of some past thoughts, and some of today’s.

Yesterday’s article has received a bit of attention from at least one high-ranking Scrum “official”. I’ve not had direct communication with them yet, but what I’ve heard reminds me of a quote from the Paul Newman movie *The Life and Times of Judge Roy Bean”:

If he wanted a chance, he should have gone somewhere else.

Truth is, Scrum has had its chance. My remarks yesterday, and in most of the articles I’ll refer to below, are primarily addressed to the Scrum Alliance, but they certainly apply to Scrum.org as well.

A moment of “fairness”.

I do not wish to condemn these organizations unfairly. I generally like to seem fair, and this is no exception. Both the Scrum Alliance and Scrum.org do pay some attention to the developer. Scrum.org has a “Professional Scrum Developer” rating, PSD-1. I do not know whether there is a PSD-2. If there is, I’ve found no evidence of it in a cursory search. You can buy a PSD-1 assessment for $200. The assessment is a written on-line examination.

The Scrum Alliance has, or is about to have, at least two levels of “Certified Scrum Developer”, CSD and A-CSD, if I’m not mistaken about the names. A number of trainers and coaches have been collaborating on learning objectives for these certifications, and I’m told that there is some definition around courses leading to these certifications. I’ve heard from a few sources that the new CSD course will not require actual programming. It used to. I have seen some of the learning objectives, and they do require the applicant to have actual ability to do things like refactor legacy code and such. It is not clear how they’ll determine that the applicant actually has those abilities. In the past, the instructor had to officially declare that the students could do the things. I don’t know if that’s the case for the revised certificates or not.

Numbers? Well, Scrum.org has published theirs. They have almost 400,000 “Professional ScrumMaster-1” members. They have less than 14,000 “Professional Scrum Developer-`” members. That’s less than 4 percent. Since there’s more than one developer per ScrumMaster, it should be more like 400 percent. The fraction is upside down.

I was unable to find similar figures from the Scrum Alliance. I’ll wager right here that the same inversion is present.

So, yes, both Scrum Alliance and Scrum.org do pay actual attention to developers, offering learning materials and assessments. And they’ve reached only a tiny fraction of the developers who need help.

It’s something. It’s not enough.

Enough “fairness”?

No, the truth is that I am trying throughout these articles to be entirely fair, to present the reality on the ground as I see it, to present what I know of what the Scrum purveyors are doing, to express why it’s not enough, why it’s the wrong strategy, and even to say what I believe the right strategy is.

I don’t know how to be more fair than that.

Let me sum up my position here, and then I’ll point out some more articles you might want to review.

Official RonJeffries.com Position

Iterative Incremental Development, as specified by Scrum, requires special skills not commonly taught but absolutely necessary, and these skills have not been effectively provided by the Scrum purveyors.

Iterative Incremental Development
Scrum is an example of an iterative, incremental approach to software development. To really succeed at this, your team needs to deliver a “shippable increment” of integrated working software in every Sprint, from day 1 to day N.
Requires Special Skills
Although iterative incremental development can deliver more value faster than other approaches, it does require special skills, including on-going testing, incremental creation of infrastructure, continual changing of code while keeping it running. Two of these skills are commonly called “Test Driven Development” and “Refactoring”.
Not Commonly Taught
These skills are not commonly taught in any of the existing forms of software development learning, whether in universities, coding schools, books and articles, reddit, or in the streets, where so many developers “learn” how to program.
But Absolutely Necessary
If the team can’t start with no code and smoothly grow the product, keeping it actually working, from zero lines of code to thousands, they simply cannot keep up with the demanding drumbeat of Scrum, which demands more and better sofware every week, forever.
And Not Provided
The numbers are clear. With the best of will, there’s no valid argument that can be made that Scrum’s purveyors have provided Scrum’s buyers, the corporations of the world, with what they need to succeed. There should be five or ten times as many real Scrum developers as there are real ScrumMasters or Product Owners. And there aren’t. The average ScumMaster has never met a CSD or PSD, much less actually having one on their team. And they don’t need one: every developer needs the skills in order to thrive.

Iterative Incremental Development, as specified by Scrum, requires special skills not commonly taught but absolutely necessary, and these skills have not been effectively provided by the Scrum purveyors.

Official Ron Jeffries Proposed Solution

Some idiot has said that you can’t raise a problem unless you have a solution, which is perhaps the least intelligent thing ever said, but I have no solution for that. I do have a proposed solution for the Scrum v Dev problem and here it is, again.

As I said, it seems like just yesterday:

The Scrum (purveyors), if they actually cared about joyful, prosperous, sustainable workplaces, would do two things:

The Increment
First, they would focus every possible training, every possible support element, every possible communication on the central notion of successful product development with Scrum, producing the Increment. This would produce demand for the ability to do so.
Nearly Free Training and Support
Second, they would get behind the creation of incredibly low-cost, high volume training and support for software developers working in Scrum. This would fill the demand.

I believe it has to be this way because, in my view, the “market” has spoken, and what it said was “Hell, no, I’m not going to spend thousands of dollars per developer to teach them to program. They’re supposed to know how to do that.”

I believe that no amount of explanation can explain to a non-programmer what programming even is, much less explain the difference between what you have to know to do it well rather than poorly. You might just as well explain flight to a snail or juggling to a turtle. So I expect that no amount of selling to corporations will cause them to invest the kind of money that Scrum Alliance and Scrum.org training currently cost–and those trainings barely scratch the surface of the learning that is needed.

I conclude that the cost in money and time of learning how to do iterative incremental development must approach zero, and that the story needs to be told to developers, not to budgeting managers. The cost in time and money needs to be so low that developers can, and will, enter into the learning path with their own time and money.

So my proposed solution is a big change in strategy, funded by the Scrum purveyors, aimed at making the necessary learning truly available to all the millions of developers who need it.

It’s a big change in strategy, but it doesn’t take much money or many people. It would require the Scrum purveyors to care, to learn, and to do what can work, or to fail while trying as hard as they can.

I’ve addressed most of my material to the Scrum Alliance, because of my long association with that organization and its leadership, but Scrum.org faces the same issues.

It would be amazingly good to see the two organizations come together for a joint developer support approach. There’s no need to compete here: it’s a tiny fraction of either of their business, and if and when they manage to start to reach hundreds of thousands or millions of developers, there’ll be a return for everyone.

OK, that’s too idealistic even for me. But it would be amazingly good.

Problem Stated, Solution Proposed

OK, I’ve again stated the problem, and again proposed a solution. Now for a quick set of pointers to additional articles of interest. I’m just listing my own, but if readers want to propose other related material, I’ll gladly consider reviewing the material and linking it here if it seems right.

For most everything on this topic, there is of course the Dark Scrum category. Here are some articles I’d like to call out specifically:

If the Scrum Alliance Cared
Yesterday’s screed, very much aligned with today’s. Sets out the problem and solution.
Down on Scrum
Bit of a screed for sure. Addresses my sincere belief that Scrum, as it is often done, actively harms people, developers in particular.
An Open Letter
An open letter to Scrum Alliance, Scrum.org, and the Scrum Community to at least find out how many Scrum-using organizations are thriving, how many are struggling. They still don’t have this information.
Responsibility - for our non-students?
Here, I suggest that we bear responsibility, not just for our own students, but for the people who are somehow trying to follow Scrum or Agile entirely without training, just surfing on the low residual waves of what we started. Unpopular opinion, but I believe it.

Enough, Already

That should be enough for this morning. I had been planning to program, but a zillion text messages last night caused me to write this instead.

Problem stated. Solution proposed. Ball in purveyors’ court.

What will they do?