On Sun, Jun 19, 2011 at 7:10 PM, Barbie <bar...@missbarbell.co.uk> wrote:
> On Sat, Jun 18, 2011 at 08:16:40AM +0300, Gabor Szabo wrote:
>>
>> IMHO while the 20-80 rule is known in many situations 4-96 is a bit
>> unhealthy.
>
> Why?
>
> For a project like CPAN Testers there are always likely to be more
> occasional testers than dedicated ones. Many of the top ten testers have
> multiple machines, so that 4-96 calculation is a bit misleading.
>
> Plus I have no problem with occasional testers, their input is just as
> valuable as those who are able to run automated smokers, as they run
> tests often on real live environments, and can spot potential problems
> that automated smokers may not.
Oh, what I'd like to encourage is to have more people setup their
CPAN.pm or whatever client they use to send out tests.
Getting more automated smokers might be interesting but
I think it would be important to encourage more test reporting on
system-perls or other perls that were not specifically built for
smoke-testing.
So I'd like to encourage more people to send in test as they
install modules.
For that encouraging companies and counting numbers of tests
might not be a good idea after all.
>> all kinds of agreements I signed like not using the
>> companies equipment for other things.
>
> This is certainly true for a number of companies, but some are
> supportive of testing environments.
So maybe the way is just further education and encouragement.
>> Maybe creating some logos of the CPAN Testers (probably even including
>> a counter) that can be displayed on web sites?
>
> We already have the smoking onion logo. If someone wants to work with
> that to create some nice logos to be used on websites that support CT,
> then I'd be happy to provide a page for them.
That would be nice
regards
Gabor
