Doug,

Thank you so much for your well-thought-out response in this thread. 

The Perl Magpie team--at the moment--is composed of Scott Baker and myself; we 
were riffing on IRC about how to provide more consistent access to test 
results, and the idea for the Magpie running in parallel to the existing system 
as a source of test results was born.  There's been some prototyping done, with 
some happy successes, but it is by no means a production-ready product.  Scott 
had shown off those results in the IRC channel, and there were a few folks who 
made suggestions, and a few offers of assistance. That's the history, as it 
stands now.

As Mithaldu pointed out in his message, a lot of the motivation for the Magpie 
was amplified by your total radio silence for the last several months. Please 
note that I'm not placing any judgement on that silence, or the reasons for 
it--I get it; life happens, real day jobs happen, burnout and discouragement 
happen.  Whatever your reasons, the fact is, you just weren't responsive, amid 
increasing stability problems with the application.

Scott and I spoke last night (US time) after we had read your message, and 
we've decided to suspend work on the Magpie, in favor of putting our efforts 
into collaboration with you and other interested folks.  I will be at PTS also, 
and I would appreciate a "brain dump" from you on how things are currently set 
up and working, and the path forward.  I'll be talking to Scott a lot (9 hour 
time difference, so in the evenings, Leipzig-time), and passing on any bits to 
him that he can help with. 

In our conversation last night, we focused on three things we'd both really 
like to work with you on.  We *aren't* looking to "take over" the project, or 
anything like that; as you noted, competition isn't your playbook, and it's not 
in ours, either.  We want to see what we can do to relieve you of the 
burdensome parts of the project that you've been managing to do virtually alone 
for years now, and free you up to spend whatever time you can and want to 
spend, doing the parts of it you *do* want to.

1.  Stability

This is the core problem right now, and one that (we hope) Scott will be 
well-equipped to help with--his day job is in Linux system administration. His 
external monitoring has been tracking uptime for cpantesters for some time now, 
and he wants to help with some internal monitoring and notification tools, that 
will possibly help expose what's causing the outages, and fix things that need 
fixing.

2.  Performance

As you noted, there are a number of Gordian Knots that need unraveling, and 
looking over the code in the github repo, I can put attention on several things 
that may help performance in the longer-term.  A lot of searches are just 
terribly slow, it's true. I believe that our experiments with the Magpie may 
help provide insight into ways to speed things up; I'd love to spend time with 
you brainstorming and refining those ideas, so that we can get them implemented!

3.  Legacy features that are missing, or need modernization

This needs little explanation, but missing APIs and statistics reporting would 
be among the things at the top of the list here.

In conclusion, we're here for you, Doug.  You've asked and asked until you got 
tired of asking and stopped--but now we're here, and we want to help.  We've 
wanted to, for months now, and we've said so, in IRC, in emails, and on Slack.  
It's not just the two of us, either, but a number of other folks have spoken up 
on the IRC channel, with offers to assist large and small.  I'm looking forward 
to seeing you in May, and hopefully getting the access and knowledge we need to 
help you with this important bit of the Perl ecosystem.

--Ruth

Reply via email to