I’ve been doing modelling for quite a while now. (For USA readers, this is British Understatement, and really means ‘I’ve been modelling for ages and ages’). And I’ve been teaching modellers for almost as long. And every time – no exceptions – that I teach UML or BPMN modelling, I learn something new. Maybe a new way to write a Use Case for some problem I’ve not seen before, or a new way of thinking about processes and process models.
I also occasionally get asked ‘how long will it take my team to get good at this’, and the answer here is ‘it depends on how good you to want to be – I’m still learning’. And this is for a modeling language which has been mostly unchanged since 1997.
So when a new modeling language comes along, it doesn’t make me happy. I’m looking at 10 more years of trying to understand what it can do, what it can’t do, and how to use it sensibly.
You can therefore understand my reluctance to start using Archimate.
Firstly, it’s almost the same size as UML, in terms of the ideas and modeling constructs.
With UML most people didn’t use it all. A Business Analyst might only use 4 or 5 models to say what they needed.
With Archimate, to create my Enterprise Architecture, I seem to need lots and lots and LOTS of new model constructs.
Also, lots of the Archimate ideas are very similar to the ones I know and love in UML. This seems to be deliberate, and a good thing. But quite a few others take old-style, pre-UML woolly thinking – like ‘Business Function’, and ‘Business Service’, and put them center stage in an Enterprise Architecture. Looking around the interweb, there seems to be a raging debate about what these ideas mean, which isn’t reassuring. Compare this with the UML idea of a Use Case. Invented substantially by one person – Ivar Jacobsen – and with a nice simple definition. Teaching people what a good use case looks like is fairly simple. But what does a good Archimate ‘Business Function’ look like?
But anyway, my customers now want to know about Archimate, and so I guess I’m going to have stick at it and learn some new stuff.
I’m starting the the Open Group’s specification, as that seems to be the source of all subsequent thinking, and it’s proving clear, but rather brief, and with few examples. I’m also looking at the ArchInsurance example (trying to find a link for this…), but again, not really enough explanation.
I also bought Gerben Wierda’s “Mastering Archimate’ book, which I have to say is quite the worst modeling book I’ve ever come across. It’s a real ‘look how clever I am’ book, talking as it does about Wittgenstein on page 1, then diving straight into the Archimate meta-model. Few examples, little reasoning about why he decides to model things in one way rather than another. Avoid.
[edit] As Gerban has been such a good sport about this posting, and has helped me to understand more about his intent in writing the above book, I’ve spent some more time on his blog, and I must say it’s excellent. Ok, quite complicated in places, but hey, this guy is an expert, but there are lots of interesting articles. So have a look at masteringarchimate.com
Anyway, back to the books and the interweb. If anyone knows of any GOOD books or articles about Archimate, please let me know.
I’m going to need them.