Skip navigation

I attended Software Crafstmanship North America in Chicago this weekend and came away feeling much differently than I thought I would. I tweeted a summary of my thoughts last night:

@dmosher: I find it interesting that the talks I agreed with most at #scna were predominantly against established “agile” principles … 😐

Twitter is a horrible medium for expressing ideas that are packed with meaning; this post is intended to unpack my tweet and experience at SCNA 2011 by focusing on the 3 talks I found the most valuable and how they formed a cohesive and powerful narrative in my mind.

Suitability vs Capability

Gary Bernhardt gave a wonderful talk entitled “Expansion & Contraction”. I think he should have titled it “Suitability vs Capability” but it was a brilliant talk and the one I considered the best of the conference. Here’s my best attempt at paraphrasing it:

Programming Language and Technology go through a constant ebb and flow of expanding and contracting over time. During expansion, these solutions are “Capability” solutions; that is, they are capable of solving problems but they are not yet suitable. Eventually there is a contraction that happens and “Suitability” solutions emerge.

Java first emerged as a Capability Solution during an expansion in the post C/C++ era. Time passed and a contraction occurred in which Java matured into a Suitability Solution for developing software.

A few times during his talk Gary mentioned that JavaScript and NodeJS are currently in the realm of Capability Solutions. At first this rubbed me the wrong way but I think that’s because I wasn’t really listening to what he was saying objectively. NodeJS is absolutely in the realm of Capability and not Suitability at the moment, but that doesn’t mean you can’t create something useful with it.

Thinking more about the heart of Gary’s talk I find myself glad that it was on Day 1 because I found that it framed my thinking for the rest of the conference.

Programmer Anarchy

Agile best practices are often framed as things we “must do” in order to be successful but I think it would behoove us to frame them in terms of suitability and capability.

Fred George gave a talk in the breakout room on Day 2 entitled “Programmer Anarchy”. His ideas might seem radical to some but I think they are an appropriate response based on an evaluation of suitability/capability for his specific situation. It’s important to understand that the following decisions derive from a company that is continually investing in new ways to make money by building very small applications. Here’s the gist:

Agile prescribes many best practices like Kanban, Scrum, XP, Pairing, Continuous Integration, Unit Testing… the list goes on.

In the waterfall era the power to determine how systems get built lies squarely with the customer. Up-front design docs and requirements specifications dictate how developers should build a system to achieve success. Through waterfall, systems are often built completely wrong. As a result customer trust is lower, developer happiness decreases, and success is not realized.

In the agile era the power to determine how systems get built is shared between the customer and developers but there is a gap in trust because of the past failures of waterfall. Agile is an attempt to bridge the trust gap by building things faster and with less bugs. Through agile, systems are also built wrong but they fail faster which mitigates risk and decreases the cost of change.

In the era of Programmer Anarchy the power to determine how systems are built lie directly with developers.

  • non-developer roles like architects, project manager, scrum-master, team lead, delivery lead, hr people and tester are completely eliminated from teams
  • developers are empowered to build software using whatever technology they choose and with whatever tools they choose
  • applications and systems are extremely small (100-300 lines of code) so unit-tests, acceptance tests, and continuous integration are eliminated entirely
  • teams are encouraged to write and re-write applications using whatever they want, including capability solutions (Clojure, NodeJS, Cassandra, Hadoop, etc..)

Fred talked a lot about his company and how they have a lot of trust in their developers to be able to turn ideas into money. The first thing they do on any project is write the business metrics code to be able to verify whether something they produce can be tied back to business value. This is an absolutely critical point that needs to be understood. Most of the time, my dissatisfaction in the work I do is related to not knowing whether something I build will actually have value. If I had metric and monitoring code in place that was able to tell me what I created is making money and can be considered successful my job satisfaction would increase greatly.

Unit tests weren’t suitable for Fred’s teams because they were writing such small applications, so they eliminated them. Traditional “Agile” roles weren’t suitable because he empowered developers to make decisions on what to use, how to use it and which developers would be best suited to building it, so they eliminated them. I think you can see the pattern.

I found it interesting that Fred’s teams got rid of things many in the Agile movement consider to be not only suitable but essential to building good software. It was also fascinating that he actively pushed his teams to use capability solutions like Clojure, NodeJS, Cassandra and Hadoop, which proved wildly successful.

Whether you agree with the decisions or not is irrelevant; Fred’s team made a judgement call about the suitability of all the typical “Agile” best practices and cut away all the things that weren’t suitable. Good teams find ways to eliminate waste by evaluating whether the things they do are capable and suitable and eliminating things that aren’t.

Propaganda, Indoctrination, Fanbois, and Education

I believe there to be a significant difference between a “Big ‘A’ Agile” and “Little ‘a’ agile”. The religion of “Agile” is often touted as the one single way to build software. The number of roles and processes prescribed by “Agile” is excessive, however, this is due to a failure to evaluate things in terms of whether they are suitability solutions or capability solutions for a given project/team.

It was probably surprising to people in the software craftsmanship movement that Zed Shaw was invited to speak, but his talk on “Propaganda, Indoctrination, Fanbois, and Education” was the most thought provoking talk at SCNA 2011. Here’s my translation:

Anyone who tells you that they have found the “one true way” to build software is a con-artist trying to sell you something.

If you aren’t writing code, you aren’t a programmer. Programmers build software, everything else is just marketing spin.

Indoctrination happens when you’re convinced to think something is the only way. Education gives you options and lets you make choices. Don’t be indoctrinated, be educated.

Agile and all the related buzzwordy ideas can be boiled down to these 3 simple ideas:

  • Make a list of stuff to do.
  • Do that stuff. 
  • Automate the heck out of everything.

(note: these things don’t have to be done in any particular order)

While he’s abrasive and rubs a lot of people the wrong way, Zed Shaw has done a lot of good for the software craftsmanship movement. He’s trying to put the focus back on learning programming instead of “Super XP Double Pairing Kanban Scrum Sauce” and all the rest of the “Agile” marketing spin. His “Learn Code the Hard Way” series is wildly successful and is actually teaching people how to program. Somebody at the conference asked him:

“So what do you think about Unit Testing and Continuous Integration and all that stuff then?”

His response was pretty down to earth:

I’ve worked for big consulting companies doing every one of those things in the past. I’ve written tests, done TDD, used pivotal tracker blah-de-bloo or whatever you’re using. Those things aren’t bad but anyone who is trying to tell you those things are the “one true way” to build software is a con-artist trying to sell you something. If you believe that stuff you’ve got a mind-virus. You don’t want a mind-virus.

I used to change my thinking drastically all the time. I’d go to conferences and experience a lot of hype and get infected with a mind-virus that led me to adopt ideas that I would later on discard. This isn’t a healthy thing to do. Nothing I experienced at SCNA 2011 was radical enough to make me change my thinking drastically, but what I did experience will help me to be much more objective about the things I let influence me going forward.

Where do we go from here?

It’s important to be objective. It’s important to question the value of things we do, especially when we do them because somebody has told us they will solve all our problems.

To me, the most important narrative of SCNA 2011, and the thoughts behind my tweet are these:

  • Don’t buy into all the propaganda, don’t get indoctrinated. Get educated.
  • Learn to objectively evaluate the tools, processes, and ideas we use in terms of suitability and capability
  • Let go of suitability solutions if they don’t give you any value. 
  • Embrace capability solutions because you might be surprised by their value.
Advertisements

2 Trackbacks/Pingbacks

  1. […] find that my good friend David Mosher has the best blog post on what was shared at […]

  2. By SCNA 2011的主题——软件匠艺 | chainding on 16 Jan 2012 at 12:26 am

    […] Mosher参加了Gary Bernhardt名为“扩张和收缩(Expansion & Contraction)”的主题演讲: 我认为他应该把标题改为“适用性 vs 功能(Suitability vs […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: