TiSCL logo, to home page







Forum: Ideas, discussion

The questions below have developed from a list which Neil Thompson originally input to the first Software Testing Retreat in the UK, in the hope of stimulating discussion and perhaps new insights. Some of these are queries to which we feel we ought to have answers, but in reality there always seem to be insufficient reliable metrics and data. Or the "answers" turn out to be surprisingly subjective and debatable.

Responses so far below are identified by the initials of the responder, which in some cases are anonymous feedback via this website (and TR means Testing Retreat generally). More responses are still invited (or you may ask additional questions!). Please see email address and notes at foot of this page.

Click on the questions to see responses so far:

What motivates software testers; why do they want to specialise in testing?
NT: They do seem to have a community and shared purpose. They claim to understand commercial pressures, but (for example) as users of commercially-sold software they tend to question why "that" had to be released before it worked properly. What are the ethics here: shareholder value or user value? How can we properly define "works properly" in terms of risk-benefit balance?

Are testers artists, craftsmen or scientists?
NT: The literature includes books claiming the first two (Myers 1979 for art, and Marick 1995 & Jorgensen 1995 for craft), but no-one seems yet ready to claim it as a science. No reason why we shouldn't reach that position, in my opinion. It can also involve philosophy and psychology. Software testing may be considered to have layers analogous to atomic physics (unit testing), chemistry (integration), mechanics (system testing) and biology (acceptance). Cem Kaner (2004) has aired a possible analogy with social science. 

Isabel Evans devoted an entire keynote speech to a question similar to the above, at EuroSTAR 2003 in Amsterdam, Netherlands. She began with the alternatives of engineers, architects or craftsmen, querying also whether testers are "professionals" (or should be). Her title was "What is success for a tester?", and she had solicited opinions from a number of testers, and others in IS/IT, as input to the talk. Alternative labels included charlatans, bumbling amateurs, failed developers... Some emphasised the need for commercial understanding. The conclusion was that testers should be "professionals" with a mixture of the skills of researchers/journalists, scientists, architects, engineers, craftsmen, technicians, entrepreneurs and anarchists!

CS: Every job needs me to build recognition by others that testing is a skill, before the true worth of the testing team is appreciated. Perhaps one reason is inexperienced test managers (and their recruitment agents) who think they need a team of developers, with testers needing to know only that programming language and not needing generic testing skills. Another reason may be that technical support for testing is too often under-planned, then the test manager tries to compensate later by panic recruitment. Often testers are prevented from involvement until too late, then are removed too early.

Why haven't they got testing all clear and predictable yet?
NT: Testing is still a surprisingly immature discipline, even though the first testing conference was back in 1972. The proceedings were published as a book in 1973 ("Program Test Methods", ed. William C. Hetzel), pre-dating Myers' "Software Reliability Principles & Practices" by three years and "The Art of Software Testing" by six years. Boris Beizer started to write "Software Testing Techniques" in 1978. But even now, part of the fun is that it is still a mixture of art and science, plus a sprinkling of other approaches to taste. There is so much still to be done, yet the ground keeps shifting to keep us on our toes (object orientation, web, the y2k "anticlimax", etc).

CS: Testing seems rarely taught as a subject in universities, or even treated properly as an adjunct to development. ISEB etc may not reach a large enough audience, so need to educate people sooner in their careers.

How much does testing really depend on testers having different psyches from developers (eg destructive v. creative)?

CS: The "us and them'" division between developers, testers and users still seems to exist (in the UK at least). It takes a lot of patience to break this tradition down, by showing respect for the developer or user and gaining their respect in return.

Other questions (updates are coming soon, as a by-product of the Retreat minutes):

If you would like to enter into discussions on these or other relevant topics, please email your responses and/or additional questions to NeilT@TiSCL.com, stating whether you want a summary of your missives posting here, and if so whether or not anonymously!

[Home]     [Top of this page]

© Thompson information Systems Consulting Ltd 1998-2004