2 ms·
Bring in as many testers as will be useful, as soon as they will be useful. In my case, I have a codebase which meets the "minimum feature set" criterion of "i
by cperciva 19y ago
Bring in as many testers as will be useful, as soon as they will be useful.
In my case, I have a codebase which meets the "minimum feature set" criterion of "is this something people would like to use"; but I've only invited in a relatively small number of testers, because having lots of people test my code right now would (a) result in me spending time creating accounts and explaining things to people instead of writing code, and (b) result in people being misled about certain key aspects of the functionality.
At the point that I bring in lots of testers, I'll still have issues of accounting (not a problem, since at least the first wave of beta testers won't have to pay), scalability (not a problem, since the beta testers will be a small enough group that I won't need as much scalability as I will ultimately have), and performance edge-cases (which will be documented as "these things don't work very well; don't worry, they'll be much faster soon") left to address; but overall I want my code to be in a state where the vast majority of beta testers say "wow, this is amazing" and run off to tell their friends about it as soon as I let them in.