On Wednesday 03 September 2008 17:02:16 Barbie wrote: > > Is the purpose of CPAN Testers to report whether a distribution will > > install on a given machine or whether its tests pass on a given platform?
> In strict terms, No and Yes, but that is only a small part of the > picture. No, it's the whole picture. If the purpose of CPAN Testers is to report whether a distribution will install on a given machine, then it's *primarily* useful to end-users. It's *secondarily* useful to the distribution's maintainers if they care about the particular machines on which it may or may not install, but start with the primary question and figure out who cares about the answer the most, and there you have the primary audience. If the purpose of CPAN Testers is to report whether a distribution's tests pass on a given platform, then it's *primarily* useful to distribution maintainers. It's *secondarily* useful to the distribution's potential users, if they can correlate passes and failures to the details of their platforms. Now before everyone jumps on me and says "You're on drugs, you hate all users, and you can't possibly believe that there's any way the mere thought that purpose one has ever considered the idea of allowing its shadow to cross our minds!", I ask one more question. What do the tests and reports to which I've objected *actually* test and report? I don't mean to say or imply that CPA Testers cannot (or has not) identified legitimate problems outside of its primary scope, but I believe that conflating those two purposes leads to confusion over which features to add and how to identify and when to report failures. -- c
