The DevOps avowed intent is to bring business, development, test and operations into a tight collaboration to deliver working software regularly and continuously. But if you want to get the business interested, you’d better choose a better name. It’s all in the message.
When I was in my twenties, I met an undertaker called Mrs D’ath. Then a few months later, a policeman known as PC Sergeant. Sometimes it seems that people follow careers based on their names and develop a great talent later. Did the name or the talent come first, Usain Bolt?
You could believe in synchronicity, or you could believe that the name was an important aspect of how people see themselves. Ask a marketer how important the name is, and they’ll tell you how the Rolls Royce Silver Mist never sold in Germany because mist in German means something unsavoury that you normally flush down the toilet.
After all, “Your brand is not what you say it is, it’s what THEY say it is,” says Marty Newmeier, of The Brand Gap, and the discerning Germans couldn’t see past the name to the quality of the car.
So it is with DevOps. The name is so screamingly technical and bland that any head-screwed-on non-tech person will fail to register it as a thing at all. As far as creativity goes, the childish concatenation of two technical terms strikes me as the first thing that somebody thought of while they were absent mindedly picking fluff from their belly button.
Yet apart from the sheer obviousness, there are three reasons the name DevOps is terrible if you are using it to drive change in your organisation.
• The name may neutralize or even negate the good work great engineers do because, “It’s just about nerds!”
• The name fails to generate any natural interest and forces you to spend more time and effort educating your audience. “I’ve not got time to talk about that propeller stuff, can we do it later?”
• You’ll face all the same challenges every time you try to introduce a new department or leader to the concept. You’ll run out of energy before any changes have happened. “What did you say again?”
Yet it can’t be denied that finding a great brand name for DevOps is a tough challenge. Google have their own version known as Site Reliability Engineering (Yawn!) and a straw poll of my work colleagues generated not a single serious contender, (though one cynical suggestion was, “Fat Raise for the CIO”).
But I suggest that a committed investment is made by the DevOps community to come up with a better name. Think about it. You want product managers, business executives, stakeholders, financial gurus and strategists to endorse a fundamental change to how they engage with technology, but you’re telling them it’s totally boring and they shouldn’t bother listening.
Better to promise excitement and status and at least have a chance of achieving magic, even if your promises are overblown. Put a grey tarnish on an old brass lamp and most business people won’t pick it off the floor to find the genii.
By William Knight
Then when he told you he was going to invite your spouse, extended family, the council representative and your neighbours, you’d probably send him packing in his ute and put a request for a new contractor out on neighbourly. It seems incredible that anyone would actively promote such an approach. Doesn’t it?
The death-rattle of a stage-gated, procedure-heavy dinosaur
But this is what happens during SAFE increment planning days. Not only are the build teams invited but anybody with the slightest interest is asked to put down their tools and spend two days developing an implementation plan. Think of the cost; all those contractors on hourly rates, the things that don’t get done, and all the expensive technicians sitting around a table discussing options. It’s enough to give finance a heart attack.
The agile community certainly reserves some criticism for SAFE. “From a Scrum perspective, SAFE might be dismissed as the death-rattle of an essentially stage-gated and procedure-heavy dinosaur. It can be viewed as a reflex to agility, and not a species of it; a last-gasp relic from a prehistoric waterfall world,” writes Ian Mitchell for DZone.
Any yet, despite my fear the increment planning would yield no workable results, in fact my first planning day showed just how effective such a program wide gathering can be.
But not in the way you might expect.
First, it was clear the plans the build teams created were unlikely to be realistic. Far too many impediments existed at the time, and it only took the first sprint after planning to see that too much had been unknown, and that the teams had not understood the significance of a whole programme increment plan – this was the first after all.
Under intense scrutiny to deliver
Yet it was the way the business stakeholders and executives reacted to the discussed risks that was the great success. Despite some trepidation, teams were mostly open and honest about the blocks to progress. It seemed that the leadership were not used to, and had not been confronted with such clear impacts before. The executive listened, took action and took actions away.
And how could it be otherwise? The leadership attends SAFE planning because of methodology commitments, and must then seriously consider the messages it receives and has no room for prevarication in such a public forum. They are under intense scrutiny to deliver.
Perhaps none of this would be necessary if business stakeholders and executives spent the time to inspect and adapt the work their build teams are actually doing – as you might expect them to do. But we all know the challenges faced getting stakeholders in regular attendance at reviews and reacting to actions coming out of retrospectives.
This then seems to be the power of SAFE; to give senior management a framework and an excuse to get involved. Without the pull of methodology, they are often unable to work directly for the build teams and without the glare of public acknowledgment are let off the hook when inconvenient actions are demanded.
And there’s another reason SAFE is a good idea in legacy-bound, large-scale organisations.
“At scale, we're still in a waterfall world and a poorly implemented one at that. In fact, I'd say that most organizations demonstrate a stage-gated culture, and not the true application of any sort of process at all. That's why they look to SAFE. The barrier to entry is comparatively low,” writes Mitchell.
“The Scaled Agile Framework appeals to large corporations. It has currency in the boardroom…” he adds.
So next time the builder suggests a planning day, you might want to take him seriously. If nothing else, it could tell you from where the hidden costs might come and determine the risk in your ideas.
Likewise for SAFE. The planning day will probably not deliver a perfect vision of the future, but run well, it should reveal all the warts, carbuncles and scabs obscuring the view. Focus on those items. Worry less that the actual plans will be wrong, and focus on revealing all the impediments.
Agilists should never be suspicious of revealing ugliness, it gives us focus for improvements and a transparent reason to wield a big mop and bucket.
Image: Matt Frederick http://flickr.com/photos/34684195@N00
Good technical skills not only help you take pride in your work and produce error free software, they can also boost the team’s happiness. Let’s take a look at just one aspect: CI or Continuous Integration.
It’s funny how things work out. Last week I posted a blog on how agile philosophies can promote a team’s well-being because they align so well with the things that make most people happy. And then over the weekend I saw that Dave Fecak writing for DZone (a famous Developer EZine) had conducted a survey on developer happiness.
There’s a mine of good information in Dave’s article and I recommend you read it. But since CI (Continuous Integration) has become a bedrock of technical practice – and hence a first level agile technique – I asked for a few of the source statistics and kindly received a reply.
Five hundred and four DZone readers** responded to the Continuous Integration question and of those 280 did CI, and 224 didn’t.
Of the 280 that did, 67 were happy, 150 were average happiness, and 63 were unhappy.
But, of the 224 that didn't do CI, 23 were happy, 111 average, and 90 unhappy.
“So the number of unhappy [people] in the non-CI shop were almost 4 times the amount of happy,” Dave Fecak says. “Thus, my theory that CI doesn't make developers happy, but the absence of CI might make them unhappy.”
Why might this be so? Using the SCARF model of Status, Certainty, Autonomy, Relatedness and Fairness, here are a few potential reasons:
I don’t think Dave Fecak’s results are an accident. Good technical practice is essential for mastery of software delivery and CI is a critical technique. Done well it improves quality, efficiency and communication which enable teams to work better, collaborate better and resolve issues faster.
Of course CI keeps you happy.
**The audience for this survey was DZone readers mostly or people that read about it on Twitter. “It's not a perfect study by any means, but I thought the results were interesting,” says Dave.
Specification by Example (SBE) requires clear instructions and time for practice for even the most basic concepts. It is not a common-sense solution, it requires a multi-month journey that starts with small, considered steps, most likely in the training room.
A picture paints a thousand words, so they say. But when you are trying to explain technical things so that business people and geeks share the same understanding, it seems a thousand block diagrams and a million process maps do not a picture make.
Which is why SBE (specification by example) is becoming so prevalent among certain sections of the Wellington IT industry. Giving business people the chance to show their thinking by writing examples seems like such a simple idea, you wonder why it wasn’t always done like that. It’s a colourful dream.
Yet black and white reality kicks in soon after teams start out. Thinking in examples is such an alien concept to most that it takes months of practice, review and experience to optimise the process. And unfortunately in these cash-strapped times, giving teams space for training and contemplation is not fashionable.
But teams will respond beautifully even late on, when given a safe environment to explore. “It reinforced the concepts of the whole agile process,” said one business analyst, after a morning’s workshop, adding that SBE, “Accentuated the importance of communication.”
And the folly of late training was exposed by another attendee who had been working with SBE for eight months without having the chance to understand first. For her, the lost opportunity was exasperating: “We totally should have done this eight months ago,” she said.
Forcing teams to take on any technical practice without them clearly knowing the importance is a crazy waste of money. In my experience, SBE can lift a team from quality duffers to zero defects in about four months from a standing start, if adequately trained and mentored. That’s not much to ask given the usual effort spent fixing things after release and the liability of supporting broken features that nobody really wanted anyway.
Practice does make perfect. If you want your teams to excel in specification by example you must give them the foundations on which they can build. They won’t of course become experts overnight, but at least they will begin the journey facing in the right direction
It's an experiment, but to promote the upcoming Agile Blog in, I have created this little video explaining how I use the Six Ps to create and finish a narrative style blog post.
To explain, the Wellington Agile Blog-in meetup, is a dedicated time set aside to write a blog have a drink and some crisps. Too many of us want to write blogs but find it difficult to get started, let alone finish. But Blogging is so very important these days, whether it's for self promotion or so that your team can explain itself in confluence to your stakeholders.
So the blog-in is a way to instil just a little bit of discipline while meeting new people and having some fun.
You can find out more at the meetup site: Wellington Blog In
I hope to see you there.
Now I know I’m not alone in being a bit crap at starting blogs. When pondering this issue with colleagues, we came to the conclusion that time put aside for blogging once a month might be a solution.
So the Wellington Agile Blog-In was born. A monthly, or so, meetup for would-be bloggers to get together and actually produce a blog, in the time it takes to eat tea and watch two episodes of Coronation Street.
On the night, we’ll dive straight into the blog pool deep end. As an ex-journalist with credentials from The Guardian and other broadsheets (quick bit of self-promotion so you know I have at least some experience) I’ll do a rapid talk on how I approach blogs and discuss the six Ps of blogging: Purpose, Plan, Produce, Polish, Publish, Promote
We’ll have a glass of wine, a few crisps and we’ll get down to blogging with a little help from your friends.
Please join the group here and come along to the first meeting on Thursday 26th November at the offices of IntegrationQA – they are kindly lending us a room (details if you follow the link and join the meetup, or you can email me at firstname.lastname@example.org)
At the end I’ll facilitate a brief discussion on the aspects of blogging we’d like to learn more about, so I can go away and find some inspirational speakers and people for the next few sessions.
I hope you can join me.
Details: Thursday 26th November – 5.30pm. Go to meetup for more: http://www.meetup.com/Wellington-Agile-Transformation-Blog-In/