There's a lot made of software testing at the moment, but some people still choose to overlook this facet of programming retail. It's important to think of right now because there are so many forms of programmed media being released into the marketplace at such a rapid rate, thanks to the insatiable demand for apps. That's not to mention the less disposable forms of software, such as operating systems, which are needed to power countless technological devices.
The counterarguments for thorough testing are that it's slow, it costs money, and not doing it doesn't necessarily translate to immediate problems for the product. Given the extent of testing that needs to be done on code to really confirm its viability as a product it can take three times as long to do as it does to code. This is obviously a problem if you're racing against competitors all the time. Not every code error means the system won't work - in fact there can be plenty of coding mistakes within a piece of software than still does everything it's supposed to. It might have subtle problems that never really materialise to the user.
But these aren't valid reasons if your product is in any way meant to be used seriously. Stress tests and alpha testing are all there to make sure that once this software leaves the lab and goes into the real world it runs effectively. If for some reason it doesn't, word will get around very quickly indeed. An example is Samsung's Ice Cream phone's 'ghost volume' issue. This only materialised when the CPU was under a lot of strain, but that was enough of a reason for retailers to take it off the shelves. Consumers don't give your product a second chance after that.
The amount of mistakes in code is staggering. While you might be able to get away with a few bugs, it's a very risky probability game to ignore investigating for problems in the first place. There can be twenty errors in code for every one hundred lines of functions, so that's a lot of glitch potential.
So software testers are necessary, and with the vast amounts of software being produced, their numbers continue to swell.
The counterarguments for thorough testing are that it's slow, it costs money, and not doing it doesn't necessarily translate to immediate problems for the product. Given the extent of testing that needs to be done on code to really confirm its viability as a product it can take three times as long to do as it does to code. This is obviously a problem if you're racing against competitors all the time. Not every code error means the system won't work - in fact there can be plenty of coding mistakes within a piece of software than still does everything it's supposed to. It might have subtle problems that never really materialise to the user.
But these aren't valid reasons if your product is in any way meant to be used seriously. Stress tests and alpha testing are all there to make sure that once this software leaves the lab and goes into the real world it runs effectively. If for some reason it doesn't, word will get around very quickly indeed. An example is Samsung's Ice Cream phone's 'ghost volume' issue. This only materialised when the CPU was under a lot of strain, but that was enough of a reason for retailers to take it off the shelves. Consumers don't give your product a second chance after that.
The amount of mistakes in code is staggering. While you might be able to get away with a few bugs, it's a very risky probability game to ignore investigating for problems in the first place. There can be twenty errors in code for every one hundred lines of functions, so that's a lot of glitch potential.
So software testers are necessary, and with the vast amounts of software being produced, their numbers continue to swell.
About the Author:
Iseb intermediate software testing is the best certification for a software tester.
No comments:
Post a Comment