The heat is on for Chris Williams, he only thought Dave Cantrell was chasing his heels, but Andreas König has also stepped up his game too. It has been a staggering month for the sheer volume of reports. In fact halfway through the month I did a quick check to see the current state of testing, and discovered that after only 15 days of testing in August, there nearly as many reports as there were for the whole of July! Andreas was less than 1000 reports behind Chris and Dave was less than 1500 behind Chris. However, Chris certainly rose to the challenge.
The use of virtual machines and space on spare boxes by some of the testers has helped to give coverage to a lot of platforms that weren't seeing much exposure. In part this is one of the reasons I created the CPAN Testers Statistics site in the first place, so I'm glad that worked. The Win32 platform has also been making a very significant return to form over the last 6 months as a regular testing platform, which extremely welcome. All the major platforms are well represented these days, even Irix. But more are always welcomed.
Again over the last 6 months, there has been a dramatic increase in the range of perls getting tested, once again largely thanks largely to Chris, David and Andreas. while some testers use virtual environments, which is something that PITA will eventually be in a good position to take even further, it's also interesting to see failures where the default install of Perl is not the one being used for testing. It's highlighted some bad assumptions by authors and has helped to improve some of code on CPAN.
Although there are a significant number of address still unmapped, I have managed to resolve 16 new tester address mappings and 7 new address mappings for known testers this month. However, there is a new challenger who has Chris, Dave and Andreas in his sights, and that's Johannes Plunien. Johannes only started testing last month and has already settled into 4th place for the monthly listings. It is great to see new testers getting involved.
At this year's YAPC::Europe in Vienna, the CPAN Testers BOF proved very fruitful, if only for a photo opportunity for Andreas, Chris and Dave. There were a few additional thoughts discussed, including the change to the subject line of emails. This is something that David Golden recently suggested to Adam Foxson and myself, and although we all agree on the change, the details still need to be clarified. Also discussed was the move towards a non-email based method of fueling CPAN Testers. Currently the email/nntp method is the primary announcement and report notification repository. However, Andreas has already been looking at an alternative method of announcing the recent uploads from PAUSE and providing a last week, last month, last quarter and last year list should not be too difficult to achieve.
For sometime now there has been discussions about the method of reporting results. There has also been the concern from some authors that they may want to receive reports for particular distributions, for only FAIL reports, or not at all. The initial discussion centred around the use of the META.yml, which although might work, requires parsing being added to the all the tools. However, using a HTTP method of recording the results and then using the central server to send authors the email, it may be more appropriate to allow authors to register to the service (via PAUSE IDs) and alter their settings. At the moment most of this is still in the discussion stage, but it will be a major change to the way CPAN Testers currently works. We still expect the email mailing list, and thus the nntp group, to still exist, as they can be useful for general discussions regarding specific failure reports, but the aim will be to reduce the amount of mail that authors receive, which are not applicable or not wanted.
After photographing Andreas, Chris, Dave and Lars Dieckow before the BOF, I've also thought about expanding the Stats Leader Board. If you meet me at future events, introduce yourself and I'll add you to the role call :)
Re:Some notes on the website
barbie on 2007-09-03T08:36:00
Unicode is all my fault. Can't blame Unicode there. I've just fixed it on my box, and will upload the new files in a moment.
Regards the ratio of results, that could be quite an interesting graph. I'll have a look at that a little later, and see about adding that to the list. Thanks for the suggestion.
