Also, more testing does not automatically lead to more bug fixing. :d
More testing should at least lead to more "bug finding" which hopefully will also result into more "bug fixing". Otherwise, if bugs that are found are not fixed, it doesn't make sense to search for them or to report them.
Now, if more testing does not lead do any more "bug finding", then this might be a sign that we have a release candidate.
However, I personally think that a software that is not a "consumer software" but rather a business software can't be run on the premise "release often / release early". Especially for research, where you often have a short time window for the field time, where you need to run your survey a major bug like "No export possible" or to "delete questions" is a no go.
I agree that smaller bugs can appear anywhere, but when I use a survey software, during the field time those things can't happen. This is impossible. Because how do I explain to a client why their survey didn't work and we had to run it again, or why we have a 2 week delay, because we have the survey results right there in front of us, but we can't export them to start analysis?
This is why I personally DON'T use Limesurvey for client projects yet. I just can't justify the risks involved. For smaller projects that are less important and critical I might use it, but for big important projects? This won't fly. I can't afford to loose clients over this.
I think my philosophy of "Stable Software" is more in line with Jelo's. A stable software should not contain any major bugs. It should be reliable and if there are bugs, they should be minor and no show stoppers. Can it happen that there might be a major bug in a stable version somewhere? yes, but the chance should be going towards zero.