Wednesday, November 15, 2017

Live from Potsdam, Day 2 - Agile Testing Days 2017

Wednesday, November 15 dawned... well, not really. It was pretty foggy when the sun rose, so I'm not sure there actually WAS a dawn. However, it did get lighter. Conference attendees and participants made their way to the morning activities of Lean Coffee, a morning run and yoga (yeah, I know I did not mention the run OR the yoga in yesterday's post - mostly because I do not do either.)

There is a full slate of activities for today - keynotes, track talks, workshops - AND! Games! TESTING GAMES! Great fun.

Jose is finishing morning announcements and the keynote from Claudio Perrone is about to start - Popcorn Flow.

Interesting title slide - with "Change is hard, make it continuous!"

Here we go!

Operational excellence is not enough, Most ideas for software ideas (and others) are really... crap. Too many times we end up with crap ideas getting delivered really, really fast - because we get told that faster is better. Innovation is key - so you get crap delivered by jet engines or crap on a stick (with spikes)

So - we want to improve, Except Improvement, real improvement, takes change. and CHANGE is a bit like Godzilla. So, don't worry about Godzilla - he's pretty big. Look at single cell organisms, they are changing ALL THE TIME.

Popcorn Flow is an anti-fragile philosophy -

Popcorn Flow Principle 1 - If change is hard, make it continuous;
Popcorn Flow Principle 2 -Everyone is entitled to their own opinion, but a shared opinion is a fact;
Popcorn Flow Principle 3 -

OK - it is too dark in here to really type... so changed seats and now I have a bit more light - maybe I can keep up

Problems & Observations - What is getting in the way of innovation (or anything else)?
To beat inertia, everyone will see problems - when multiple people share the view that something IS a problem, it goes from an opinion to being a FACT. There is a shared view of an issue that is supporting inertia and getting in the way of not delivering crap.

You can form shared observations to create and elicit options (get a bunch of people, like 3 together and explore the ideas around. Experiment - like - "We might try BDD or Code Reviews or Paired worked to maybe make things better..."
    - pull quote! If there is salt in the cake, it is already too late! We can't fix it!

We can talk and agree on experiments, look at their expectation then TRY them! What is the result?

Set an expectation - then see what happens . Not meeting the expectation is not a failure - it is a correction of the anticipation.Be prepared for that form of failure - but - learn from each one.

Fail fast is not the point - LEARN fast is the goal!

The thing about "being agile" vs "doing agile" is not something to worry about -

(I perhaps celebrated a bit too much with my friend Huib... sheesh)

If there is a continuous flow of experiments (the core of agile) then what limitations are there? Some may work really well - KEEP DOING THEM - some may not work - STOP those, try something else.

Try "Here's a problem as I see it and some ideas that might help..." with the manager - then experiment with the experiment model.

DING! Quality is something we worry about - what about the value we get from something? Like, a conference talk? Does a "high quality" talk deliver value to EVERYONE? Or is value to the customer without regard of the "quality" of the presentation - or software (ahem).

Set yourself up for learning - Do something - oberve what the results are (for you and your customers) Is there a correlation between the 2? There may be, but not always - so look for the gaps.

And he slides into an example - His (then) 5 year old son had an idea - he could sell snails!

First few experiments failed - welp - give up? What about trying something else?.So, instead of giving up after the 1st 2 tries - he launched a 3rd idea - a comic book about... snails. Launched - landing page - AND... collected 1 Euro per donation, and send updates with pdfs in the future.

First night - 24 Euro. and it just kept going up - 1 in 6 conversion ratio (that is pretty astounding)

 Don't be afraid to think big - really - like - end world poverty...

Keep thinking - keep innovating.

RIGHT - sat down and helped a speaker prepare for her session (SOME one changed all her slides.... last night) I know, I still have not loaded the images from the keynote - I'll get to them a bit later - I promise.

And HERE we go!

Wearing Hermione's Hat: Narratology for Testers - with Marianne Duijst.

Narratology is the study of stories - the narrative and other aspects of any story. (warning academic theory stuff coming)

The Fabula is the series of related events - logically or chronologically - the "plot". There there is the Story - the presentation - ANY presentation, the manifestation and colouring of the fabula - the plot.
The TEXT is the medium these are presented to the oberserver or reader -

Building blocks to consider -
The AUTHOR (the one who write the narrative) who is likely NOT the narrator - they are not the same person (usually).

So - J K Rowling presents the story, but the Narrator herself is unnamed.
We experience the actions thru the experience of the characters (Ron, Hermione and Harry) - THEY will experience things based on the different perspectives.

For example - 1st quidditch match - Harry nearly falls off his broom, etc., and we see the story thru Harry's perspectives - however Hermione sees something else.- the change in perspective.

Hermione's Role is to understand and figuring out the fabula / plot /events of the story...
HOW? She gathers and interprets information from MULTIPLE sources - multiple stories. She stores this information and retains it for use.

This is usually not obvious to people who read the story once - usually people read the story once and leave it. Multiple readings will shed more light and reveal more information as we can get past the mechanics.

This works for software as well! Code? how often do we look at code? Maybe once? What are we missing? (We never presume we miss anything - just as we don't believe we miss anything when reading a story - once.

The question is - can we work around this?

Who do we trust?
Rita Skeeter does this weird thing of talking with people - and twisting anything she gets into something else. Hermione, by observation, realizes how Rita works - and finds a way to present an alternate story to the same facts - the same detail.

She - Hermione - listens to all sources - including those that appear to be minor r insignificant - like, House Elves. She realzies that elves such as Doby and Creature have extremely powerful magic - magic stronger than hers. She also is able to retain and present this to others because she gained the understanding needed.

This can set up shifts in understanding from the various perspectives - we change the realities - based on this shifting perspectives. The change in paradigm follows this all - for example - Snape kills Dumbledore in book 6 - and is slagged as irredeemably evil - until we see the reading of Snape's memory.

We are actively NOT looking for all the pieces of the information that can be used - because they do not meet OUR narrative - what things do we hold that we are not considering deeply. Our closely held beliefs/stories /understanding can we wrong. We don't like that. It is uncomfortable.

Context drives everything - Time is part of the context - as is place - The Time Turner allows for shifts in time to allow us to see the actions Place is another context. We know what we know because of where we are.

The challenge to Harry - which Hermione takes on - is to enable others - where Hermione teaches Harry about the information needed - aqnd shows Harry, Ron and others on how they can do the same thing.

The problem is, Hermione does not always have all the information needed to make the best decisions. She builds a model based on incomplete, if not faulty, information. Just as we do with software.

For software people, we need to find ways to think differently - what happens if we do this - what can we infer from THIS. What happens if we change...


Lunch and many excellent conversations. More on those later -

Now it time for the 2nd keynote of the day - A Practical Guide to Testing in DevOps, being pesented by Katrina Clokie - the AWESOME Katrina Clokie.

When she was young, she had a teacher named Miss Perfect (really, that was her name) who taught the kids in the class about Atoms - and how they are everywhere and make up everything - including how they could move atoms around by moving their hands in the air - which is made up of... Atoms...

So, when DevOps began to be a thing, many testers looked at that and said "Oh, that is for Devs and Ops folks, we're not in there..." Except... well, Atoms

In Agile, we have everyone helping with testing and helping with quality.  And then we have others added in the mix of Agile to come back as .... Dev Ops - so Infrastructure, Support, Analytics - everything, no?

Citing the equally awesome Dan Ashby. his model - where testing is involved everywhere...


Once upon a time.... Katrina's org was having a lot of issues. Specifically big red lights going off when builds were made - and things went wrong regularly.  Machines dying and stuff happening and conflicts and - ick.

By splitting the Selenium Grid from the Jenkins Slave, ...
Things got better - - a LOT better.

Now, while this was not a typical "DevOps item, the fact was, that Testers led the drive to make things better. Awesome!

They also test in production. Ummm - yeah - Did I mention they are a bank. Testing in production? Really? Dangerous. Not really.

Monitoring. watching - looking for behaviors that might inform us on what is actually happening.

Because - in a mature DevOps environment, the testing is all around - at every level. Because testing is not looking for bugs - it is looking to see how the organization can be informewd about the bahvior of the software - before and after it goes to production. Really.
And a post to the LeanPub site! YEAH!


Sliding over to Getting Ready for the Cloud with Sabina Simons. Sabina is a very bright, intelligent young woman from the Kitchner-Waterloo area of Ontario - practically a neighbor of mine! Only around 4 hour drive and an international border crossing away.

And we're underway - OH YEAH!

She is a Development Manager at D2L Corporation. She was given an option of 3 choices for next position in the company - 2 of which she had already done, and the third (the one this story is about "scared the crap" out of her...

What does "the cloud" mean? Welp - their background was to move to an AWS single tenant instance - fine - why?
They wanted to be able to move faster and serve customer needs more quickly and focus on getting product out to people and not the logistics of how to do that. Their customers are educational institutions at all levels, from pre-school to universities.

Can they get this to work? Locally? Globally? How does this actually WORK? What in the world does this look like?>

Change infrastructure - except they have a heap of a mess of legacy... stuff.many, many, many levels of ickiness (my word - family friendly blog)

Some 20 streams of work present, their code base was comparable to... a park with trash everywhere. No test instances, not test cases, not automation of any sort. With a mindset of getting it sorted - their goal was stated as "leaving it a little cleaner" each time they did something. Addressing immediate legacy needs by making incremental improvements.

Experiments! Canary Testing - small actionable instances, small changes to track what is being done. Leave benchmarks and small unit tests documented in the process - and set changes up and run things in the perf lab... things are awesome! Except the code path they were interested was not executed.

Complex scenarios sometimes take unusual solutions - like, how to set things up and how the conditions exist in production. Really.

Benchmarks are fine, making them drive forward. The set the throttle and exercise it.
Things worked great! Except - when they opened it up - stuff hit the fan.

So, now they have a significant problem. Loads of people pointing fingers and loads of "YOU SCREWED UP!" So, deal with the situation as it is - look at what process flow can be tweaked to improve the situation and open conversation. Stepping away from the blame game and being open in communication can actually help calm everyone down and make things better.

Home grown automation tools - (looks like some instrumental graphs she is displaying - and yeah, they were were instrumental stuff - which is awesome - I get to ask a question!)

Tracking behavior through various models - good solid reliable information. Watch what is happening and look for leaks - Really look in the corners, the places that appear to be OK - particularly when things are ... a little odd.

Use the code analyzer to track what is being used - presumptions & beliefs are grreat - BUT - go prove it.

Failure mode testing -

Always sounds great until people realize what that may mean. In short - How do you figure out how the system will behave in different "suboptimal" ways - usually catastrophic!

These start with What If? stuff - and working through it. Sabina's team worked through the issues as a team - (YEAH!) the question is always figuring out the stuff that is likely to happen.

Enter the Master of Disaster - writing alerts to handle with some level of grace things that are likely to go wrong.

Takeaways -
Lead by example. Move boldly and show actual results;
Experiment - What Happens if I Do BLAH!;
No Excuses (really) - everyone has stuff that needs to get done, make things happen;
Share your story - internally, externally - make the world a better place....


Right - I snuck off to a corner for a few minutes quite time, then sat with some new speakers and had a lovely chat with them on talking in public and working up the guts to stand up and bare their souls in a way that many people don't realize novice speakers do.

I then did some preparation for the Agile Games night, where I was running a table of mixed small games confusing people and offering puzzles that required them to set aside their preconceptions and immediate beliefs and look at the "call of the question" and the words actually said and written.

Great fun for everyone. I gave away many stickers for prizes and people had fun. There is some socializing going on right now that I intend to return to - with a "tester cabaret" which seems to be great fun. IN the meantime, I've had excellent conversations today and tonight which have given me much food for thought. More on that later.

No comments:

Post a Comment