26 August 2011

Cardiff STC Meetup Writeup

I attended my third Software Testing Club Meetup last night, this one being held in Cardiff. As is usual for these events, I spoke to a lot of interesting people not just testers but recruiters and developers also. I had a conversation about automation is the big thing that companies are looking for, to a developers about testability and test driven development, as well as a number of other software testing activities (e.g. documentation, performance etc).

There were four lightning talks given, the first given by my friend and organiser of the event, Sean Robbins. He showed us how to use Watir to run an automation test (in this case, search for something on google and check for results) by using it to run the browser, or by just sending HTTP requests which allows the test to run much faster. Very useful if you do not need to touch the form, which can be very susceptible to breaking your automation test from the smallest of changes.

The second was presented by a developer, Warren Seymour, who was talking about testability in code. He expanded on the concept of MVC (Model View Controller), which normally is only used serverside, to show how you can use it for client side code as well. He recommened two javascript MVC frameworks, the cunningly titled Javascript MVC and Backbone.js.

Thirdly, there was a quick overview on performance testing requirements from Vicky Dibble. She made the very reasonable point that a performance requirement cannot be simply 'the page will load in under five seconds', since that is giving no information on user load or considering the simple fact that sometimes, especially with web pages, you just get a slow response without a seeming underlying reason (hence she suggested changing the requirement to '98% of the time, the pages will load in under five seconds'). What she showed that was of particular note, was to use the User Community Modelling Language (UCML) to build up a picture of your systems typical user flows and be able to work out what tests you need to create.

Lastly, Mark Coakley talked about the issues facing him where they have lots of very old legacy systems that require regression testing, but they are also hiring a lot of new people from outside the business and have to be able to get them up to speed as quickly as possible. What he showed was how they began testing their documentation looking for clairty and accuracy. They then build up a picture for each document with a red, amber, green lisiting for each area they are looking at, which indicates that it is completely wrong (red), wrong but usable (amber) or all ok (green).

I highly recommend these meetups, so be sure to check the Software Testing Club Meetup page for upcoming events and do be sure to attend one. I'm hoping there will be another in Bristol in the next few months and will be sure to ramble on about it here when it does.

15 August 2011

An Agile Problem

  I have myself a problem with the proponents of Agile. Before I state it, I'll need to clarify a few points:

1) I have read the Agile Manifesto and support the values and ideas behind it

2) I have never worked in an Agile environment (or even an iterative development style project)

3) I very much consider myself as a context-driven tester

  With these points in mind, let me share my issue.
  Through the things I have read from those who do work with and contribute to the Agile methodologies, I get the impression that Agile is always right, and if it doesn't work, it must be how it was implemented (see we tried baseball and it didn't work). Therefore, there cannot be any context where Agile may not be suitable. This sounds suspiciously like best practice to me.
  However, I freely admit I may be entirely wrong here. As I say, it is an impression based on various things I have read and from a position of relative ignorance. I would like this impression to be changed and certainly to be less ignorant on the subject.
   Links to suitable reading matter would be appreciated.

11 August 2011


  Thanks to my newly aquired twitter account I noticed a tweet by Matt Heusser throwing open the possibilty of an Association for Software Testing Conference (CAST) to be held in Europe. At the moment EuroCAST is only a fledging idea, but it is a very good one. I urge people to show support this idea, via the usual social media or any other way you wish to do so, so that we may have a EuroCAST 2012.
  I also think we should get it held in Bristol, UK, for the sole reason that is where I live.

P.S. EuroCAST is not to be confused with the Gold and Silver finished charms from http://www.eurocast.co.uk/

09 August 2011

Value of Testing

  A couple of years ago, my company had to make some redundancies and we went from essentially a team of 5 testers to a team of 2. I was under no illusion that I would survive a second round of redundancies if they became necessary, having heard many times that testing is the first to go. That second round was announced a couple of weeks ago. 
  I am very thankful, and more than a bit gratified, to discover that not only were neither of the testers going to be made redundant, but that we weren't even required to be part of the consultations and so were never even considered being made redundant. I like to think that this proves that the work we do is not only valuable, but recognised as such throughout the company.
  Having seen and taken part in many discussions about how you show the value of testing, the moral I take from this is simple. Good testing shows its own value.