Our Blogs  


Eating our own dogfood

By Dave Snowden  ·  June 6, 2014  ·  Musings, News

Back up to Bangor for the second time in a week, bad planning that but it was unavoidable.  This time to pick Tony up from the station before driving on to luck in The Caban and then an afternoon working with a new software team in their converted (or rather converting) Chapel. I've shared one picture here and there are more on flickr.  We were there to run through the current functionality of SenseMaker® Explorer.   We will be launching iOS Explorer shortly; long hold ups as we had to rewrite for the current version of iOS and shift the coding resource in house.  I'll write at greater length on our plans for a three level replacement for Explorer in a few weeks time.

Interesting, in parallel with this meeting I was getting some inclination of the marketing machine that is SAFe on my own blog.  Someone called Jessica started to peddle a marketing message.  I started to suspect when the odd direct comment, nay insult did not throw her off message.  Then direct questions as to commercial interest and identify received evasive answers to the point where after due warning I cut her off as SPAM.  I have many objections to SAFe, but one (and relevant to this post) is its attempt to claim outcome based management under conditions of uncertainty.  Now I have also criticised SCRUM as suffering from premature convergence but that is a different criticism.  SCRUM has huge value when the convergence of need has reached the transitionary stage between complex and complicated.  Its also authentic to AGILE principles although not co-terminal with AGILE, a common mistake.  One of the big things we are doing in Cognitive Edge is to adopt tools and methods designed for complexity as pre-SCRUM processes within an AGILE framework.  I'm finishing off an article at the weekend that will outline some of that thinking in the wider context of project management.  It will be in Sunday's blog post with a link to the draft article for anyone interested.

Now in a small company with limited resources you simply don't have the time to go through multiple iterative trials.  You need find ways to co-evolve, using different expertise without over prescription.  That served us well with the first version of Explorer where Cynthia was given a very loose brief and mutual trust got out a project that could not have been specified and continues to have high utility to this day, albeit with problems on scale.  The same was true with iOS Collector; download SenseMaker® from the App store or the Android equivalent and use activity code AgileAU2013 if you want to play with it.  There we had some patterns of past use, some ideas about the future but we trusted to developer to add functionality as they worked.

I know the value of this from my own time building decision support systems.  A lot of my best work came from ideas while I was coding; the context created understanding and the fact I had built trust with the users gave me the right level of freedom to both innovate but also to make (and recover) from some errors.   What matters is that you have a rigorous architecture and the right level of testable deliveries.  Now for some years in Cognitive Edge's history I allowed that understanding to move outside my immediate management control and I think we suffered as a result.  For a long period we were over ambitious and I was guilty of believing what I was told about deliverables and capabilities.  Either way, lesson learnt and Peter is now completing an outstanding set of work to give us a rigorous base and we are now moving to some rapid delivery of the user front ends.   But as I say, more on that later.

What is key to this approach is that you trust the people. not just the process.   I like the way the Agile manifesto assumes that the team knows what is best, but I suspect that applies to the situations where the requirements are better know and more resource is available that we do.   In that context trust has to be in people as well as process and that takes time to build.   I've been working with Liam in connection with Bangor University for over a year now and I know he understand at a deep level what we are trying to achieve.  I also love his set up in an old chapel at the foot of Elidir Fawr.  With his team they are converting it into a pirate ship, we met in the bridge.  I had heard about the chapel but this was my first visit.  Entering I smelt the old smell that I remember form school days when we were marched crocodile fashion to the Bethesda Chapel for indoctrination.  Then I saw an eclectic mix of musical instruments, living space, tools and technology.  Several of the team grew up in the area and have know each other since school.  All of that felt good as had the conversations over lunch.

As we talked through the context and need they started to add in ideas that we had not thought of.  Always a good sign.  We also shifted them away from wanting precise definitions.  I suspect that is an IT defence mechanism against users more than capable of denying earlier claims.  So I left feeling confident that we can make rapid process.  Given the work we have already done the effort to get an Explorer replacement is not high, but I am also looking for long term development here including stuff as yet only nascent in character.   I committed to spending time with the team every few weeks, although I may take the odd side trip up Elidir Fawr during a pre-SCRUM mini sprint or two!  Peter is talking about moving his caravan up there as well so we should get some intensive time being spent.  I'll be taking notes as well.  I have used this approach several times over the decade but I need to formalise and situate it as a method.