and this[^]

My native languageis German. Many years back I worked on a Siemens mainframe which used a German OS. As it was command line based (GUIs where unknown then) all the commands were in German. For example the command 'edit ' became 'ediere '. It was irritating because so many commands were not real German verbs but modifications of them to make them into commands. It simply felt that the German language wasn't really fit to be used on a command line.

the one-based day is for the VB programmers. The rest of us get the other one.

1. Does testing by third party will uncover all the errors in the program No. Nothing uncovers all the errors in the program. If you were to rephrase the question to 'is testing by third party more effective?' then the answer is: 'only if your in-house testing skills are lower than that of the third party'. So look at your skill sets and those of the third party. 2) Does Regression testing helps in removing the undetected errors I think you want to know if regression testing will help find defects that were missed in system testing. The answer is 'somewhat'. The prime focus of regression testing is to have an economic way to re-assure that the current release didn't break bits in areas other than where you made changes. Regression tests tend to be test cases that had been run in the original system testing. So if regression test cases find bugs it means your maintenance changes had a side effect that you were not aware of. Is this an error that was undetected in system testing? A matter of argument but besides the point. 3) Does Automated tools are the ideal choice in the place of manual testing? No. Automated tools have their place but they are no replacement of good manual testing. Too many people try to do with manual testing stuff that is better done with automated testing. But inversely there is much that is not well testable in an automated fashion. Specifically in system testing manual testing is more and more effective. This is a long topic so I'll just make a few assertions: - manual testing discovers more bugs than automated because manual testers are more flexible - building a robust automated test case takes about 10 times the effort required to do the test case manually - so it is only worth where you would run the test more than 10 times - building robust automated test cases is a Software Development task. You need to provide the skills, processes and infrastructure to support it. Otherwise you are wasting your money. Stephan

what do you want to test: that the application works or the fact that the passwords can't be decrypted?

I have started digging around a bit starting with ServicePoint and SoapRequest. However it looks like there is quite some way to go until I'd come to the point where the threads actually are being put to sleep. I suspect it would happen somewhere in the HTTPRequest class. you might be able to shortcut this as to where in the .NET space the dispatching (and supposedly putting to sleep) is likely to occur. Unfortunately there is this thing called work getting in the way. Cheers Stephan

I want to do set up automated testing of Web UIs. ASPUnit seems reasonably decent, but how do I test Javascript in web pages?

This did indeed fix the problem. However it does raise a question. Currently, when the number of threads trying to open simulatneous connections to the service exceed the limit, these threads go into WaitSleepJoin state. So far so good. But when the other threads close their connections, shouldn't these threads be activated by the system? At the moment the threads stay in their suspended state until they get killed by the main process (if they are background threads). Is this a bug or a feature?

Thanks for the info. I will check this out. At least it appears the problem is with my test code and not with the app I was testing. Stephan

Maybe I am missing it - but I am not connecting from IE. This is a class dll that is running under nunit doing web service calls. it does so by creating multiple threads that all hit the same web service. Basically some threads simply stop and never return; Stephan

I am just doing a bit of load testing of a web service. I wrote several test cases in NUNIT to fire off. One test case creates several (5) threads which all start close to each other and call the web service several times. For some reason, usually only one (sometimes 2) of the threads complete, the others seem to hang. It is also never the same thread that completes. Does anyone have a clue why this may be the case? tia stephan

benglish72 wrote: Ego sum insanus, sed sum felix. Felix sanus est, ergo insanus tam insanus non est! Is that some unix script?:)