Over a beer at Sheffield last week, Brian Swan commented that the XP early adopters – those that were originally attracted to XP already shared the values. This has lead me to muse on the values, whether I genuinely shared them, and whether they were reflected in my actions. Am I XP Correct?

During the break of Mike Hill’s Transitioning to XP tutorial last week, I realised that my problem with our team’s transition was that I had no real feedback on how it was going. Mike said “Could you ask the team” and I answered “They wouldn’t know”.

Bzzzzzzzt. I’m sorry, that was the wrong answer. Should we take a look at the team you chould have won?

That was not XP Correct of me. I fail on all counts, particularly respect and communication. Fortunately this is not a finite game, and I have learnt some things.

Xp2005 is definitely over; I’m tired. I’m sat in the (closed) bar waiting until it’s time to go and get my train.

I’m pretty sure that I learnt a lot of stuff. I promise to write some more up when I’ve more energy. I might do some lightning writing on the train.

Apart from some great stuff on Storytests, the best parts have been on the touchy feely side. Tim Bacon and Dave Hoover’s workshop had some great stuff on applying narrative therapy to team issues. Today was Kent Beck’s tutorial on mind mapping (or how to change your life and XP with big sheets of paper and felt tip pens).

I probably learnt the most from the conversations in the bar or over coffee.

Sunday Morning
The courage to communicate (soft woossie skills, etc….) – Diana Larsen

I need more of this stuff. Going to one of these sessions isn’t going to really help, but it’s a start. I am definitely going to search out some NLP training soon. Particularly after talking to Ben from WDS. They sent their entire team, on NLP training and it changed the team and the individuals.

I don’t know that the underlying NLP is correct – in fact I very much doubt that it is. But like Neils Bohr said “they say it works even if you don’t believe in it.”

Sunday Afternoon
Transitioning to XP – Mike Hill
Mike is energetic, charasmatic, and has a lot of fantastic insights. One is that the industry is demoralised. Amongst the thousands of terrible developers out there, many have just been crushed by years of failure and being treated as pluggable programming units. If there morale can be lifted a surprising number of these can rise to become competent social geeks.

Another insight is that XP is more a culture than a collection of practices. The practices simply establish and nourish the culture. A transition to XP is a cultural, rather than technical, change. At the end of the month is the Cynefin course in Greenwich. Cynefin, amongst other things, is about investigating and changing culture. Mike talked about the transition to XP involving a dip into chaos; Cynefin practitioners take organisations to the edge of chaos to precipitate change.

More and more I can see the relevance of Cynefin to agile. For instance I hope that the narrative techniques of Cynefin can be used not only for requirements crystallisation, but to document the culture of XP.

I don’t really have many qualms about re-editing blog posts. I’ll flesh these out later.

Monday 20th June – Keynote – Ian Sommerville: Extreme Programming for Critical Systems?

Ian Sommerville presented his view of the challenges of Critical Systems Programming, and where he felt XP fell short. It was an interesting talk, though he seemed to misunderstand aspects of XP. He particularly misunderstood the XP concept of Customer, confusing Customer with user- this is not surprising as it is quite a nebulous concept. Other areas really fell outwith the area that XP (specifically) dealt with.

Bar Panel session.
Well not really. I sat in the (closed) bar with Mike Hill, Charlie Poole, Brian Swan, and Alan Francis listening in on the chat. It was an educational experience.

Lightning Writing Workshop – Laurent Bossavit, Emmanuel Gaillot
That was fantastic. I’m quite fired up about writing now, hence the sudden rush of blog entries and revisions. The technique practices was simple – a timeboxed (10 minute) session of continues writing of anything. It’s a great technique, but I’m tired of blogging just now. More later.

PS Laurent wrote on of my favourite papers on XP, The Unbearable Lightness of XP.

Here are some quick notes from Sheffield. They are a bit hurried….

Saturday Morning – 18 June 2005
Rick Migridge – Expressing Business Rules in Fit

This tutorial was the primary reason that I came to the conference this early – I’ve been particuarly interested in “Automated Acceptance testing” recently. This is the first time that I’ve encountered the term “Storytests”: I like that it; it doesn’t overload “traditional” softare terminology.

A few months ago, I saw Storytests as tests for the system interface – functional or user tests. Recently I have seen the value of using them to express business rules and plugging them into appropriate parts of the domain. Rick’s tutorial has helped consolidate this view of the tests as primarily means of documenting user stories. I find the idea of executable documentation attractive, and early results seem to indicate that it is an easy concept to sell.

Other actions from this are:

• I am going to start trying to use Rick’s FitLibary – the Fixtures seem much more expressive than the older Fit Fixtures.
• I am going to bump Rick and Ward Cunningham’s FIT book way up my (prioritised) reading list.

Saturday Afternoon
Charlie Poole – Test Driven User Interfaces
I was not really sure whether I should attend this tutorial, particularly as it is focussed mostly at Fat clients – something I haven’t worked on for a long time. I learnt a lot of stuff, which makes it worthwhile.

Charlie’s main point was that what makes UIs hard to test is what makes other code hard to test – poor separation of concern. I possible when testing UIs, fake the model and test widgets separately – compose related controls into widgets to test separately.

It made me wonder whether I should rethink writing my own taglibs for JSP, development. Separting the model from web apps is possibly too much hard work, unless I brough Cactus back into the mix – a cactus \ Fit \ JWebunit hybrid- now there’s a thought.

In case you were wondering, the aesthetic aspects of UI can not be automatically tested (lining up of buttons, etc…).