Software Architecture Guide

Martin always has good points. What I miss from him is proofs. The weeks vs months part crossover for internal quality... How is it measured? Is it just a thought, or a wish?

Unusually useless post. An aggregator that says little. Too bad.

In the public sector of Denmark we’ve been on a decade long journey toward better Enterprise Architecture. We have national guidelines for how we want public system/services to be build and how they should be able to transfer data through APIs. We don’t get completely technical, telling you what tools you need to use to build your software, only how your software needs to fit into the environment that is public sector digitalisation.

The reason we do this is because we’ve seen what happens when we don’t.

The municipality where I work operate a pretty common amount of IT systems. We don’t know exactly how many because there isn’t a centralised unit to handle them all, but our guess is 300 different IT systems. There is a lot of common things these systems need, like authentication. That used to be terrible. Before web-services and federated IDPs it was quite literally maintaining our organisational hierarchy in 300 different IT systems manually. And that’s just our place, in Denmark we have 98 municipalities that all have 300 IT systems run by different suppliers. With a common architecture we defined how organisational data needs to be interpreted by these systems, we also build a single co-owner database where each municipality could make these data available in the same defined format, through the same APIs.

It’s already saving us millions, just from this tiny piece of the much bigger picture. It’s not been without backlash either, because as it turns out, private tech companies don’t actually want to build the smartest systems. They want to sell us silo systems, because then they can sell us access to our data 300 x, where 1x could be organisation data and another x could be statistics.

Anyway, I think we have evidence that architectural choices matter a great deal. At least when you look at it from an enterprise sized point of view.

With that said, I’m not sure I buy detailed application architecture, where you design everything in your favourite markup-language, either. We certainly have no evidence to support that this sort of architecture isn’t just a waste of time. But for the big picture, in enterprise sized organisations, you can’t live without IT-architecture if you want to be efficient.

If you want to read more about it, you can at: http://info.rammearkitektur.dk/index.php/Forside but it’s in Danish.

For some of these speakers, I’ve literally never seen any evidence that they’ve written a single line of code. Why should we listen to Fowler? What large successful projects has he created using his ideas are there that I can look at? Where’s the data behind his slides? It’s all just talk talk talk.

I’d listen to someone like Chris Lattner talk about how to design and write software because I can see whatever approach he uses must work well. I should listen to these full-time speakers because... I don’t really know and perhaps we should start challenging their talks by asking for the data behind them.

Thank you.

If I want to design excellent architecture I will crack open SICP and transpose it to the fabric of modern computing.

I also take issue with much of what Fowler et al say, but that's another conversation.