Agile Open Space: On Certifications

The AgileAustin Open Space has kicked off. I found the opening session positive, engaging, really fun to be there. I'm chuffed because I felt confident enough to suggest (convene) a few sessions. One responsibility of a convener is to capture notes from the session. This blog post is meeting that responsibility.

Y'see, one of the sessions I proposed was entitled, "Certified ScrumMaster??" I wanted to gauge people's opinions of the CSM certification: Is it valuable, is it real, or is it just résumé-padding fluff?

I got my answer. The groans, grumbles, and rolled eyes around the room confirmed my suspicions. In this crowd at least, it is worse than irrelevant. It is counter to and detrimental to the philosophy of the Agile community.

After the proceedings broke into less formal conversations, I caught up with some community members whom I respect and enjoy. They elaborated on their earlier non-verbal remarks. It is not difficult to get a CSM; you attend a two-day class. That's it. That sounds like just enough knowledge to be really dangerous.

Why do we even need certifications? What does a certificate indicate about my real skills, abilities, and past experience? No, I inherently reject any model that sets up a gate-keeper-style hierarchy to knowledge—a system that says, "We know things; you don't. Your ideas and perspectives are not as good as ours until we bless you and permit you to be one of us (and your check clears)." I don't accept religions that do this, nor governments, nor software project methodologies, for Pete's sake.

The true flaw in the CSM is the name: Certified ScrumMaster. Go to a hiring manager and ask which she'd rather have, someone with 3 years' experience as a Scrum team member, or a Certified Scrum Master [trumpet fanfare]. Those in the know, know that ScrumMaster is the role; you hear it without the space. But to those who are not yet well versed in Scrum, it sounds like Mastery of the Scrum process; they inject a space between the words.

The Training page on the Scrum Alliance website says it plain: "The journey to mastery begins with..." and "These courses [CSM and CSPO] provide a solid foundation to help you make the paradigm shift to managing a project using Scrum." [emphasis added] They state straight up that this is the starting point. But the name of the certification doesn't say that. The opportunity for misinterpretation will get people into trouble.

Why does it get my dander up? Personally, because it threatens to be Another Damn Thing I gotta do to stay in the game. Professionally, because agile projects can be magnificent, and certifications smack of the process-for-process'-sake mindset that turns software development into a tedium. Philosophically, because neophytes will incorrectly elevate the merit of opinions from a Certified ScrumMaster, no matter how little experience he may have, and dilute and muddle the tenets of Agile.

Mike Cohn the other night joked about the CSM culminating in a tattoo. I don't know, man, I might put more stock in that, if the tattoo embodied the Agile Manifesto. (Embodied—ha!) It would at least convey the right level of commitment.

Simplest Responsible Thing

Someone on Saturday asked rhetorically, how do you reconcile the agile philosophy of "do the simplest thing possible" with good OO design principles. That apparent conflict stems from misquoting the first part. The more useful version is:
Do the simplest responsible thing that works.

That is, responsible to highly probable future features; responsible to your known performance demands; responsible to your business's security and reliability requirements.

Above all, responsible to your future teammates (and your Future Self) who will have to support and maintain that code. Happily, this last responsibility is nicely supported by the emphasis on "simplest."

Note also that it is "simplest," not "easiest." Your old familiar habits will be easiest, but if those habits were formed in a different paradigm—say... old ASP that you learned by osmosis (Hey, we all had to start somewhere.)—then it will take some dedicated study to turn the responsible thing, the simplest thing, into something you can do easily. Meanwhile, the state of the art in Simple will keep moving, but that just gives us something to keep striving towards. That's what keeps it interesting, right?

Wrangling Rhinoceri

I've been asked to help some teammates become more comfortable with Rhino.Mocks. I'll practice my explanations here, and y'all can offer guidance and ask clarifying questions if you're so inspired. I'll be glad for the help.

Rhino-what? What-mocks? What-what?
The domain we're discussing is unit testing, the code that developers write to verify the logic in their code. Rhino.Mocks is not a testing framework, but a mocking framework, which means it allows you to mock (simulate) parts of your application. You use it along with a testing framework like NUnit or MbUnit.

If you have written code without keeping an eye on testability, it is likely to be witheringly difficult to apply Rhino.Mocks after the fact. Life is simpler if you write the tests first, therefore Rhino.Mocks plays well with Test-Driven Development, which drives you to design your code in a testable way.

As far as what it is, Rhino.Mocks is a dll you include with your test project files. You create a reference to it from your test project, and it gives you some additional classes and methods you can use within your unit tests.

State-based and interaction-based tests
There are two main styles of unit tests, state-based and interaction-based, and each is useful for different things. With state-based, you set up some variables, run them through the methods you want to test, and then verify that their state has become what you expect. You use Assert statements to assert, "The state should be [this]. If not, fail the test." If you're testing an addition method, you can verify the state of the answer, as in, "I assert that 2 plus 2 should be 4."

Rhino.Mocks is used in interaction-based tests. Here you are verifying the interaction between your classes. For example, in a Model-View-Controller architecture, you could test your Controller's interactions with the Model and the View, as in, "I want to verify that, when the Controller is asked to display customers, it gets customers out of the data repository, runs the results through my CustomerFormatter class (I don't know, it's just an example), and sends them to be displayed on the user interface."

You have three interactions there, so you'll probably write three tests. You are not testing what comes out of the data repository, just that the Controller talked to it. In fact, you don't even need a real data repository for this test. If you could just simulate that repository in a way that let you verify it was called... This is what Rhino.Mocks gives you.

What next?
Martin Fowler has written a great article, Mocks Aren't Stubs. If this is your first introduction to Rhino.Mocks and mocking frameworks, perhaps not everything in that article will click with you. It is still worth reading now, and I'll link to it again in my next post. It wouldn't hurt to read Martin's article twice.

My next post will include why you would want to simulate classes instead of using the real thing. I also want to explain why you even want interaction-based tests. I have an intuitive sense of the value, but I am not yet good at articulating it. Feel free to chime in.

Globally distributed scrum

I'm on record as saying that globally distributed scrum teams can work, but sub-optimally. I'd like to retract and revise that position, so that something I'd said doesn't inadvertently lead a team down a poor path.

My team's scrum adventure lasted from March 2007 to April 2008. (Why does it have an end date? Buy me a beer, and we'll talk. But none of the people who were a part of or were affected by our efforts wanted it to end. The project was a success.) We started out with four developers in Austin, six developers in Bangalore, a Test Lead in Austin, and five testers in Hyderabad. I'll call this the Two Parallel Teams period. Team members were moved over time, so that we became two developers in Austin, eight developers in Bangalore, a Test Lead in Austin, and three testers in Hyderabad. This is the One Distributed Team period.

We had a good collection of team members, and I think we could have worked together very effectively, had the distance between our desks not been so great. I am not criticizing off-shoring; I am criticizing scattering your product owners, developers, and testers across different continents. Note who I included in that list there; if your dev team is co-located, but your product owner is far away, you have a distributed team.

We did not have a choice about where our teammates should be located, but we had influence over which project methodology we would use. So the choice was not "distributed scrum versus co-located scrum," but instead "distributed scrum versus distributed waterfall."

Do I still think distributed scrum is better than distributed waterfall? Yes, although with less unbridled enthusiasm. It required more meetings from all team members (including product owners) at crazy hours, but it still made it easier for the product owners to get the system they need; it still had us writing more code, more often, which is way more fun than fighting over contracts; and it still enabled productive, visible progress toward our goals.

We were happiest in our Two Parallel Teams mode. The developers in one city worked on one feature, and likewise for the other city. The only people who really suffered here were the product owners and the scrum master, who had to accommodate a ten-and-a-half-hour time difference to have their sprint planning meetings.

We offset our two-week sprints by one week, so that the testers could be shared between the two sets of developers. It worked okay because it did not require the "save state and hand off" process inherent in distributed teams. That process is, at the end of the day, you write an email explaining what you did, where your thoughts are about the work-in-progress, what you want your teammates to pick up, and what you want them to leave alone because it's currently too fragile to explain or share. It's like "hibernating" in Windows--save a snapshot of where you're at--and it takes time.

The more distributed the team, and the more they need to collaborate, then the more time will be spent handing off, and the team will be less efficient. In our One Distributed Team mode, the hand-off process overwhelmed the development process, so that more time was spent saying what you're doing than doing. This is the dangerous mode I want to warn people away from. If multiple people are collaborating on a feature, they have to be able to talk to each other, in person, real time.

There are still dangers lurking in the Two Parallel Teams mode. The most insidious is the way it undermines egalitarian, democratic, self-organizing decision making. For a team to decide things like designs, system architecture, even working practices, they have to be able to discuss and debate as a team. This requires trust, camaraderie, and personal safety. Those things are built in informal, day-to-day interactions--y'know, the way friendships are built. Over email and conference calls, they happen incredibly slowly, if at all.

You have to be comfortable with someone to be able to argue with him. The threads stitching together a global team are tenuous, and we operate with great delicacy to avoid snapping them. The communication channel is so anemic, you have to choose every word carefully, to ensure it does not accidentally offend. This is the opposite of having the confidence to know that, after we argue about this topic, we'll still be a team. Which means a lot of decisions go un-debated; they get made by default instead of by consensus.

Do I recommend globally distributed teams? No. Regardless of the project methodology, and no matter how talented the team members are, you will be dramatically less efficient if the members of your team are scattered over the planet. If you're off-shoring, then off-shore--put developers, testers, product owners, project managers, the whole team in one place. Software development is intrinsically collaborative. Structure your teams so that they may collaborate.

This experience gave me a taste, however, of how fun scrum can be. With a co-located team, you could seriously rock and roll.

Software Development as Manufacturing Metaphor

Alistair Cockburn's talk the other night was inspiring, thought-provoking, and entertaining. If you have the opportunity to hear him speak, I highly recommend it.

I have been wary of the agile community's current enthusiasm for the Lean literature, as applied to software development. I'm hesitant to trust "manufacturing" as a metaphor for "software development," because I work for a large manufacturing firm. My executive management believe they are experts in efficient manufacturing, and therefore apply those lessons to managing software teams. The agile-to-lean trend gives them permission to make this leap.

The problem comes in attending to the wrong part of the metaphor. Incorrectly applied, it overlooks that developers are knowledge workers, and you are led to believe that they are interchangeable laborers who should specialize on a small task. Easily outsourced, among other things.

But Mr. Cockburn made me realize the right part of the metaphor to apply, the useful part. He said, substitute "unverified decisions" for "inventory," and now you can map your team's process, look for bottlenecks, and improve throughput. When unverified decisions pile up, work is not getting done.

Mr. Cockburn said that different types of bottlenecks mean you need different solutions. He also flouts the strictest adherence to Lean, which would have you remove all waste, and instead suggests spending efficiency to increase throughput. You have too many developers and too few db designers? Let the devs iterate through a bunch of designs and code until they have it really baked, before handing it over to the db designer. Too many analysts and not enough devs? Get the requirements solid, signed off, and thoroughly detailed before giving them to the devs. Too few product owners to tell you which route to build? Build both and let them choose.

Spend your excess efficiency in order to increase the speed of the overall team.

Back to my beef with the manufacturing metaphor. Incorrectly applied, it lets you think waterfall software development makes sense: They give us the specs for a computer, we build it, it gets tested. But the scale is wrong. What a waterfall project actually does is give you the entire order for the DoD's workstation upgrade initiative, and you build 1000 computers before any of them get tested. (A huge backlog of unverified decisions.) Rapid iterations bring you back to building one computer at a time. And Scrum makes the deciding and the verifying nigh simultaneous...
Developer: Purple or green?
Product Owner: Purple.
Developer: 2gigs or 4?
Product Owner: Ooh, 4!
Tester: Oops, hold it, you actually grabbed the 2.
Developer: Whew, good catch.

What I learned from Mr. Cockburn is that the Lean philosophy holds useful lessons for the agile practitioner. Drawing an analogy between writing software and installing chips on a motherboard does not help you, but thinking of your software team as a process comprising different roles, each with their own decisions to verify, can help you increase your team's productivity.