Software: Testing times for enterprises

gallery

Software testing might not be the most glamorous part of IT, but for businesses developing their own applications, it is essential. Good testing is key for both in-house applications and heavily customised, off the shelf software; Gartner calculates that the worldwide market is worth US$13 billion (£8.4 billion) a year.

However, like much of the IT industry, software testing is undergoing a period of consolidation. The recent deal by CSC, the IT integrators, to buy AppLabs is further evidence of this trend.

But should software testing be kept independent of software development and IT integration services? For IT directors it is a difficult question. On the one hand, dealing with a single supplier is easier and cheaper.

On the other hand, keeping testing separate from development should ensure that applications are analysed "warts and all" and there is no incentive for the integrator or developer either to cover up coding errors, or indeed sell additional services that might, or might not, be needed.

Research carried out by SQS, a software testing company, suggests that IT teams value independence highly. SQS does have its own axe to grind: following the AppLabs/CSC deal, it claims to be the largest independent software testing house. None the less, its research makes for interesting reading.

SQS hired consultancy PAC to study the market. They found that 86 per cent of IT managers believed that using a third-party testing company speeded up development times and eased pressure on in-house teams. A full 91 per cent of managers thought it was important for software development and testing to be independent of each other.

There is a further, practical point that the PAC research did not draw out, but will be all too familiar to IT managers: the need to have a testing and quality assurance system in place, that cuts across all software vendors and developers. Although "captive" testing teams might well be very good at what they do – they will, after all, know the software they work on inside out – few businesses buy all their software from one supplier.

Relying on developers' own teams could result in multiple teams of testers each working on one part of a project, without a clear overview of the business process as a whole. And asking one IT integrator to examine the work of another could be another management headache, unless the testing team operates independently of development, with strong Chinese walls between them.

Another option is to use independent testing tools from vendors such as CAST, or Quality Center from HP, but IT departments need to take the time to learn how to make the most of them. This might only be worth the investment, for businesses that carry out much of their development in house.

But the critical point is to make sure that software is tested, and tested again. The PAC/SQS study found that two thirds of companies measure "errors in production" in their software. The real worry must be the one third of companies, that do not.

Stephen Pritchard is a contributing editor at IT PRO.

Comments? Questions? You can email him here

Email to a friend

Print this page