The Future of Software Testing?

Well it’s Xmas, and there are a lot of programmes on the TV about the best of 2005 as well as people speculating about the market trends, house prices and economics of 2006. Who am I to leave a good band-wagon behind.......?

Part of my role as a Managing Director is to look forward and see what 'might be', where TCL could fit into it and where we can add value to our clients. For the last 6 years we have done relatively well in this regard, but I am posed by more of a challenge this year...in fact I would say that my answer is still a work in progress!

Primarily, my views are that the convergence of Companies, Commercial Propositions and subsequently systems is driving a lot of the IT change we are involved with. It seems to be a world wide, multi sector trend, and areas such as VoIP, Mobile, Data, Fixed line and ISP convergence is a good example of how things are unifying. Let me give you a highlight from BT. You can now take out a Mobile tariff and join it with an ISP, landline and VoIP bill. The systems you have at home enable you to have a wireless (and wired) LAN, and when you come home your mobile will act as your home (fixed line) phone - routing calls through VoIP or your fixed line as desired. It’s brilliant because its simple (almost seamless) for the user, it’s cost effective and it’s easy to buy (online).

The challenges we face as testers of this kind of technology include:
The size of it all - things are getting smaller. Usability requirements are very different. The scale of the interfaces and joined up thinking across the organisation though is much much bigger. Needing good solid design work to make it a reality rather than a travesty.
The Complexity - lots of things in the same place. To test it you have to know a lot more, about a lot more things
Time to Market - the competition is fiercer than it used to be and time to market drivers are more acute in a lot more of the sectors. A statistic I was given at a recent networking dinner was that a new software product is outdated/superseded or competitively out priced within 16 months of its launch. Projects cannot last a year or two in domains like that
Data - lots more data in even embedded systems. The performance of the systems handling this data and the security / integrity of the data in the system and more often now across the Internet is on everyone’s minds - although not necessarily in the design!
Price - Excessive pricing is no longer tolerated. Quality and Innovation may be differentiators but in markets such as mobile telcos where there is near customer saturation the cost mark (even for very innovative products) is essential to get right.

So how does this shape the future of testing? It would point towards a necessity for more Non Functional Testing, for better, quicker design and for testers that can do more and know more about a lot more things. The future of software testing therefore is rooted in proactive, professional development. Significant specialism in some areas and amazingly competent all-rounders for others.

The thing is though; there is a doubt in my mind generated by History. Love it or loathe it, you just can’t ignore history. You see when I first started in testing I was lucky enough to work in defence. We had a very mature product, in a very mature industry and the testing processes had been evolved over 20+ years to something akin to clockwork. Many of the key features brought up at today’s conferences were already common practise even then. Regression testing, testing of requirements, TTRM/VCRIs etc all sorted. So where's the new stuff? Where's our evolution? How are so many people working in 'testing' by doing exactly the same thing every day?

We thought that automation was the key to the future, and indeed I have heard that said many times in the last ten years. But look around and you see that circa 85% of all testing is still manual. So what happened? Were there no market drivers to encourage time to market initiatives, innovative approaches or change? Did no-one do this? Well a lot of us did, but why didnt it become a uniform approach?

And finally, we have all seen the need for better professional development within testing but after all this time there are still very few Universities who even mention it as part of a Computer Science Degree let alone Research and Develop it. Certification courses really are in their infancy and we spend a great deal of time reminding people that testers have to be able to test not just push paper. There isn’t an industry body, a globally accepted and common practise standard and there are commercial interests getting in the way of the true development of better tools, techniques and practises.

So if History has taught me anything its that Software Testing has been pretty much stagnant over the last 10 years (In a way like the surface of the earth seems flat from space) . An industry of silos for a long time and in fact the last 10 to 15 years haven’t really got us substantially further forward as a community.

This leads me to an interesting thought....
For me Software Testing will only have a different future from what we do now if there is a period of commercial consolidation and consistent governance/control of standardisation. This isn’t needed just in the UK, but worldwide. There is enough financial reward in software testing for this to start happening, and indeed some companies in software testing have now floated - generating capital to invest in such consolidation. When this starts to happen we might just see the merging of practices across industries and sectors generating a new level of collective competence and standards which will drive the way we work, the skills we need and the development of people that the industry needs to get on a par with other areas of IT, and perhaps start to lead the way.

The future of Software Testing in my eyes is in Unification and collective advancement. Let's see if 2006 proves me right.

Merry Xmas 2005 Everyone. I wish you all the best for 2006, and look forward to meeting up with many of you at StarEAST in May.
Stew