Technology selection 2012 – a sketch

I’m not suggesting this is the right way to choose your multinational’s ERP suite, but I’ve just made a technology choice. The success of my decision is yet to become apparent – I’m documenting this without the benefit of hindsight.

I hope I’m brave enough to do a follow-up post when it becomes clear whether I made the right choice.

1. Identify the constraints

I have chosen a Content Management System for a small company’s website. We may have to integrate with an app already written in Ruby on Rails, and that’s the programming language my developers are happiest in.

We’ll almost certainly deploy to Heroku, which adds additional constraints like a limited selection of databases.

  • Constraint: Ruby on Rails, specifically Rails 3+
  • Constraint: Heroku-compatible
  • Constraint: MongoDB, Postgres, MySQL ¬†and a few others

2. Draw up a shortlist

Aside: I’m starting to mistrust Google because of their evil attempts to promote Google-hosted content over other services. Boo. But Bing is no less evil so I still use Google for the moment.

I Googled my search for a Ruby on Rails CMS. I eliminated opinion pieces from 2009 and earlier. I made a list of CMSes that matched my constraints.

The list was

3. Final selection

To select a winner, I read the marketing fluff on each website. I then forgot everything I’d read and got on with making a rational decision.

I made a subjective judgement about how easy my users would find it to create & maintain content using the admin tools for each CMS. In fact they were all OK. I couldn’t rule any of them out on this basis.

I then looked at the GitHub repository for each tool. I looked at three things:

  1. How active the codebase was
  2. How many people were contributing to the project
  3. How successfully they dealt with issues that had been raised

All the projects had been updated in the last week but had already gone through several releases. All looked to be quite mature code under active development. A good sign.

All the tools had a wide range of contributors: Locomotive 29, Refinery 199, Radiant 59.

All the tools seemed to be using GitHub to manage issues, which helped speed up my research. Here there was a clear winner

Tool Open issues Closed issues Ratio
Locomotive 38 232 16%
Refinery 16 1196 1%
Radiant 41 267 15%

Admittedly a crude measure, but I like Refinery’s clear-up rate. And I even like that they have had more issues – it matches with their wider community involvement.

We’re going with Refinery, but the others are respectable choices too. If we discover a showstopper early on with Refinery I’d have no qualms trying one of the other two.


3 thoughts on “Technology selection 2012 – a sketch

  1. Hi Dom, interesting to see you making key IT decisions. I have been documenting our decision methods this year so it is interesting to see you blogging on your real world one. I was surprised you did not put benchmark the products against the overall business objectives as one of your steps….anyway hope it was a good choice


    • Hi Chris, great suggestion.

      If I was choosing tech to configure and deploy to users immediately then I think I’d automatically do this. In this case, I’m choosing a toolkit to get us a head start on a project, so the overall project objectives are probably a better benchmark.

      I guess my quick play with the admin features of the CMS tools was my equivalent of that! I should probably have thought more analytically about that exercise, and written it up as a separate decision phase, you’re right.

      • Hi Dom

        Yes indeed and for a small firm like yours this kind of choice is one piece of the jigsaw but one of the things we have learnt with our customers is that taking a step back and looking at the overall company objectives can sometimes change the choice in surprising ways. What I also found interesting in doing the strategy for some of the mdedium sized service companies was how a not so obvious overall objective for the business like say “we want to expand through acquisition” could change the choice of software….

Leave a Reply