Saturday, 17 April 2010

ACCU2010 writeup

Here’s a short (OK, I got carried away a bit) summary of the conference and the talks/presentations I attended. I only went down for one day - Friday 16th April due to work commitments.

Friday morning:

Keynote talk from Russell Winder “The Multicore Revolution - We’ve only just begun”. Originally Dan North was down to present “Simplicity – the way of the unusual architect” but couldn’t get to the UK due to the airspace being closed because of the Icelandic volcano eruption.
Russell is clearly very knowledgeable and presented a good talk about the architecture of computers the way different programming languages address those. Good presentation that was probably more interesting to the developer community than for me but interesting nonetheless.

After a coffee break which I used for frantic telephone calls and trying to book train tickets I attended Rachel Davies “Understanding User Stories”. Not having written any before this was an interesting presentation with just the right amount of practical exercises for the participants. She mentioned the book “Agile Coaching” which she wrote or co-wrote which I’ll have a look at.

She explained that she and her team used a user story template which is now being used by many people around the world. I got the impression that she didn’t expect that and was pleasantly surprised that such a small thing created to make life easier in her project team has such a big impact. She did not that this template doesn’t work in all cases – especially when there’s no discussion with the business or if technical implementations are the order of the day rather than user workflows.
The template followed a
As a...
I want...
So that...
format which goes a long way getting all the necessary information onto a small card. She noted the importance of this being a physical card rather than a spreadsheet or electronic document. To me this format is similar-ish to the way use cases are created as the actor (As a xyz..) and workflow (I want...) and goal description (so that..) are all present. That’s not bad and I don’t confuse the two but just note the similarities.

After a practical exercise of writing user stories for an imaginary tool of our choice we discussed that. What I found interesting is that breaking down the user stories to the right level and amount of detail is really a team decision – something many people struggle with when writing use cases.
The importance about the user stories should be “Do the user stories help us understand user context and business value.” That’s a harsh abbreviation and there’s more to it but to me these are the essentials (together with acceptance criteria).

At lunch I met James Bach and after listening in to another conversation we spoke about some senior managers and their viewpoints on testing, the TA model (parent/adult/child modes) and implementing rapid testing in companies. As I’m following both James and Michael B’s blogs there was nothing groundbreaking new for me but it rather reinforced my own views but also highlighted that the problems I’m facing are not with the understanding of the rapid testing approach but in other, more business related areas. I left contemplating and thinking about this (which is a good thing) but will need some time before I’ll find answer. Interesting.

After lunch I attended Paul Field’s “Change, change, change – the 5 year evolution of an agile team.” It was an interesting presentations about his experience about starting out with a 2 man agile team taking on a failing project in an IT unfriendly environment – the internal customer (another department) got burned in the past and was wary of IT in general.

He spoke about the pitfalls of building a team and scaling that from 2 to 7 after a while. Generally a good talk although it could’ve been a bit condensed from my point of view.

I liked his use of the Timeline Retrospective where he mapped colour coded issues as experienced by the team in sprint retrospectives over the whole project time. This can be a good way not to measure the way the project is going but the way the team is feeling about it which doesn’t necessarily needs to be the same. (And indeed was not, the team thought that they delivered badly at a sprint when the customer was delighted).
The retrospective starfish was also interesting consisting of 5 areas for
More of
Start doing
Stop doing
Less of
Keep doing
In the burnup chart he presented he mapped the story points “Done” and the story points as asked for from the business. This was not per sprint but over the whole project which showed that after 4 months (in a 6 month project) the team had created as many story points as the business originally wanted (great) but wouldn’t be able to finish by immovable 6 month deadline – a big problem. This burnup chart showed the problem in time and action was taken to get the project back on track.

When new team members where added he found that his former “flat” structure didn’t work anymore as existing team members were more senior in terms of knowledge so not everyone was equal. This needed to be addressed and after the new team member were brought up to speed through pair programming and other measures the team picked up some momentum.

After a coffee break I went to Astrid Byro’s “Just in time – last minute testing” presentation. She jumped in for another presenter who couldn’t come due to “Volcano awareness week” as James Bach coined it.

This was the weakest talk for me, the title didn’t have very much to do with the talk which was basically a war story about a 2 year RUP project with very high risks (accepted), lots of stress and a “command and control” approach should have died out 20 years ago. System migrators where used as sole testers which would’ve made me uncomfortable (maybe she was as well). I can see that from a PM point of view using people with domain knowledge is a good idea and using testing as a training exercise so you save handover further down the line is good for project timelines as well. But I can’t shake off the feeling that professional testers would’ve done some good – apparently that wasn’t possible (or thought necessary). At the end of the project performance testing was started (always a bad idea due to the high risks involved, again they were lucky and their application performed well). Their clients infrastructure didn’t but that wasn’t seen as a problem as the project in itself delivered, even if that didn’t actually help the customer. My view is more holistic than just pointing at the customer and saying “it’s their problem, our application performs well”, but maybe that’s just me.

That concluded the official presentations but I saw about a dozen lightning talks, the highlight for me was James Bach’s “And...also” presentation (I’m predictable in that regard, I know).His fiery talk had a lot of passion and humour to be able to appeal to even the most hardened developers and I believe he got some people thinking. It was about thinking not only what the expected positive outcome is but also what he also expects to be there (or not). For example NOT getting error messages, expecting the result to appear for more than half a second, etc – all things an automated test suite might fail to notice but a human being would notice – sometimes instinctively, sometimes with a bit of thinking.

I won’t mention the other lightning talks, there were too many to mention and I didn’t take notes of all so wouldn’t be able to give a complete picture anyway.

After a longer break the conference dinner started which was very interesting as well, after the starter we played musical chairs with only the speakers remaining seated so that people got more opportunity to speak to different speakers and other attendees. At my first table I sat next to Jeff Sutherland, creator of Scrum, without recognising him which I found extremely embarrassing. I promised to put in a good word with the conference organisers to get him back next year so that I could have a longer discussion with him – so organisers, if you’re reading this, do all of us a favour and invite Jeff for next year as well!

That’s it, thanks to Volcano Awareness Week I’ve been able to write down these notes while on my train journey rather than sitting at home already so let’s all say thanks to the volcano gods


1 comment:

  1. Hi Thomas,
    Good write up! I must have been in 2 sessions you were in! So disappointed we didn't get to meet up properly.

    Gutted I missed the lightening talks and meal.

    'Understanding User stories' was interesting in the vagueness of everyones story.

    I agree in your summary and was somewhat dismayed by the 'Just in Time testing' talk.