LabVIEW
Do you take part in LabVIEW betas?
Broken Arrow wrote:With all respect Cory, I disagree. They know exactly what they've changed, and they want to know what we think - how it's working, etc. To load a program like LabVIEW and just "look for bugs" would be a full time job IMO. Hence my lack of participation, I guess.
This is a very simplistic (and wrong) view and shows that you probably never participated in a real beta.
A software system such as LabVIEW is extremely complex and it is impossible to fully test all possible combination of scenarios. If A works and B works, we cannot conclude that A+B works equally well. Real-world use can often be different to what the developers ever imagined.
Beta testers are not instructed to simply run labview and "look for bugs". That would be silly. The best you can do is use the beta environment in your daily normal work and report when something unexpected happens. This does not require much additional time (except for keeping backups in the original version and downconverting before releasing anything in the production version, and possibly recovering from a crash once in a while).
Beta testing is a very important process in software development and I am sure there are even mathematical models for it. All software contains bugs, that's a fact of life. Beta testing cannot find all bugs (as evidenced by patch releases), but it will ensure that there are 100 times fewer bugs once the software is released, many fixes for bugs that would have never been found by NI internal testing. The number of bugs found is directly proportional to the man-hours spent using the beta. Testing is not a full-time job. Even if you would only use it on a single rainy Sunday afternoon and find one single typo in the online help, it would have been worth it. Fixing a bug during beta is orders of magntude cheaper than fixing a bug after the software has been rereased.
All NI can do is invite beta testers. 10% will find 90% of the bugs (wild guess!), some find a few bugs, and some will never even install it. Still, the number of found and fixed bugs will be roughly proportional to the number of beta testers that participate. The more the better! 🙂
Of course there is an upper limit on what can be done, because the number of qualified programmers (A hard upper limit is the population of planet earth :D) and the time duration of the beta are limited. Only if an infinite number of programmers test for an infinite amount of time, all bugs (except for the last one!) can be asymptotically eliminated.
A beta tester should be quite familiar with LabVIEW. It should not be a beginner who cannot find the diagram, reports wires turning black when not connected, or reports problems wiring two controls together. 😄
Beta testing is essential. Do you really want NI to throw a beta quality product at the consumer and simply wait for a barrage of bug reports coming in from use in mission critical applications (I just upgraded and now the nuclear reactor is stuck at 100%!)? Thought so! 😮
So, I respect your decision not to participate in betas, but please don't try to justify it with flawed and contrived arguments. Beta testing is definitely not for everyone.
When was the last time you released a program that contained no bugs ("hello world" does not count!)? What do you do to ensure a quality product?
Some useful readings:
Beta Testing (read it!)
Enjoy! 😄
