On Wednesday, 17 June 2015 at 18:35:48 UTC, Andrei Alexandrescu wrote:
> You wouldn't just not go to work for two weeks and then show up at your desk as if nothing happened. This kind of stuff needs some level of planning, barring exceptional events. For example, Adam could tell his readership "I'll be on vacation next week, so no issue on August 9". Which is totally fine.
>
> I'm very happy Martin is doing well but I am disappointed about this unprofessional behavior. In fact I could only assume the worst because he seemed one of the most serious people I've dealt with, so this came as quite a major breakage of trust.
>
> With this we revoke Martin's role as release czar. His github access will remain the same for the time being.
>
> Until a replacement is found I will fumble with the release process myself, and I could use all the help I can get. In fact it would be a nice gesture of Martin to help with the transition. I'll be also counting on help from the more process-oriented members of the community. One note - 2.068 will be delayed because I have a house move to deal with for the time being, not to mention being busy with std.allocator and std.collection.

As most others here have noted, this seems too harsh, particularly for a volunteer position.  The truth is that release manager is a role that basically nobody wants to do, as it has responsibilities but basically no compensating reward, and by removing Martin from the role, you might actually have made his life easier!  It doesn't sound like a punishment to me.  As I recall, our last release manager also disappeared for awhile, likely because he had a real job to do.

Martin probably made a mistake by not making clear that he would be away from contact for a while, but I wonder if it was ever made clear to him that a head's up was expected, particularly if a release wasn't imminent.  I understand that you would like for us "to evolve from a tribe to an organization," but I always thought that was unrealistic for a volunteer community with essentially no money coming in.  The fact is nobody is paying Martin to show up at a "release manager" desk every day.

Unless and until companies using D, like Facebook, Sociomantic, EMSI, and whoever else is using the work of this community to make money for themselves, deign to invest back some money behind functions like these- release manager could probably be done as a half-time position- they will not be done to the standards of a commercial entity.  Which is likely fine with the rest of this volunteer community, as we can't ask for what we're not willing to chip into providing either.

On Thu, 18 Jun 2015 04:21:10 +0000, IgorStepanov wrote:

> This is international forum and we may allow only one kind of intolerance: we may hate programmers, which doesn't use D yet.

we shouldn't hate them. they are poor people that simply can't see The Light yet. but even php coder can become a programmer. and programmer can become D programmer.

On Thursday, 18 June 2015 at 07:22:51 UTC, Laeeth Isharc wrote:
> Yes - one gets so used to being able to reach people in the modern world, that a short period of disconnect leads to all kinds of concerns and doubts.

Yes, it is interesting that new media creates new spaces and new types of perceived presence. In one way it is cute that people can worry about lack of presence (in the loving way) and if people feel they worried for no reason it is easy to get upset. I think everybody has experienced that one way or another.

On the other hand it basically means that if you don't pick up your cellphone you end up having to explain what you were doing at the time. So you have to fight more for your privacy. I personally find that intrusive.

So it might generally be difficult to get the _best_ person in a group to volunteer if you need high availability. It would be better to have a plan with milestones. Then you know which week participation is critical at least a month ahead.

And really, for a compiler consistent quality is critically important, so getting more of the few best skilled people to volunteer is a lot more important than getting many lesser skilled to volunteer.

The price is to make it enjoyable for skilled people to volunteer. Whatever it takes.

On Thu, 18 Jun 2015 10:09:44 +0000, Joakim wrote:

> basically nobody wants to do, as it has responsibilities but basically no compensating reward, and by removing Martin from the role, you might actually have made his life easier!  It doesn't sound like a punishment to me.

exactly. being free of that work is a blessing. especially if we consider the absense of working non-abstract roadmap (and even if there will be one, how can people be forced to work on it instead of working on what they want to?).

Andrei is passionate, but little too harsh. at least that's how i see it.

On Thursday, 18 June 2015 at 09:05:27 UTC, Marc Schütz wrote:
> On Wednesday, 17 June 2015 at 18:35:48 UTC, Andrei Alexandrescu wrote:
>> With this we revoke Martin's role as release czar. His github access will remain the same for the time being.
>
> I'm speechless.

+1

On Thursday, 18 June 2015 at 11:14:25 UTC, ketmar wrote:
> On Thu, 18 Jun 2015 04:21:10 +0000, IgorStepanov wrote:
>
>> This is international forum and we may allow only one kind of intolerance: we may hate programmers, which doesn't use D yet.
>
> we shouldn't hate them. they are poor people that simply can't see The Light yet. but even php coder can become a programmer. and programmer can become D programmer.

Well said!

On Thursday, 18 June 2015 at 10:09:45 UTC, Joakim wrote:
> Martin probably made a mistake by not making clear that he would be away from contact for a while, but I wonder if it was

I don't think it matters if Martin did or did not make a mistake. Everybody make mistakes and talk past each other. Period. That's to be expected and can be addressed in an enjoyable manner, but is harder to do in written media (finding a path where the humour in the situation overtakes over the perceived seriousness). So we need to be conscious about finding those paths to the more loose non-formal atmosphere.

What matters is that the community needs to pay the price of having skilled people volunteer. And the only payment to available is to make the task enjoyable. Most people find the more loose non-formal atmosphere enjoyable, I think.

On Thursday, 18 June 2015 at 02:22:13 UTC, Vladimir Panteleev wrote:
> I would like to ask, what can we improve in our tooling and infrastructure to lessen the burden on release czars?

Merge dmd, druntime, phobos repos.

On Wednesday, 17 June 2015 at 18:35:48 UTC, Andrei Alexandrescu wrote:
> On 6/17/15 5:38 AM, IgorStepanov wrote:
>> He returned back to github and posted some messages. End alarm:)
>
> Glad to hear that. However, the need for two other release managers is still there.
>
> The role of Release Manager and "This Week In D" are the two ones somewhat special in the community: they're closest to an actual "job", in the sense that regularity and professionalism are crucial. You need to "show up" and do it. If "This Week in D" is not there on Sunday evening, it's not a weekly. If we have no release manager and no contingency plan, we can't release. In the future we'll have folks depending on this and planning ahead for it.
>
> You wouldn't just not go to work for two weeks and then show up at your desk as if nothing happened. This kind of stuff needs some level of planning, barring exceptional events. For example, Adam could tell his readership "I'll be on vacation next week, so no issue on August 9". Which is totally fine.
>
> I'm very happy Martin is doing well but I am disappointed about this unprofessional behavior. In fact I could only assume the worst because he seemed one of the most serious people I've dealt with, so this came as quite a major breakage of trust.
>
> With this we revoke Martin's role as release czar. His github access will remain the same for the time being.
>
> Until a replacement is found I will fumble with the release process myself, and I could use all the help I can get. In fact it would be a nice gesture of Martin to help with the transition. I'll be also counting on help from the more process-oriented members of the community. One note - 2.068 will be delayed because I have a house move to deal with for the time being, not to mention being busy with std.allocator and std.collection.
>
>
> Thanks,
>
> Andrei

As others have said this seems a bit harsh.  I presume that by 'we' you mean that yourself and Walter arrived at this decision together?

If you really felt the need to remove Martin from this position, the appropriate way would have been to discuss this matter with him privately, rather than effectively firing him on a public forum.  That also comes across as somewhat unprofessional.

Based on the the discussion on the forum it appears that the manner in which this decision has been executed has effectively alienated pretty every good candidate who might have been a replacement ... in addition to almost starting an OT religious holy war.

On Thursday, 18 June 2015 at 12:37:15 UTC, Dragos Carp wrote:
> On Thursday, 18 June 2015 at 02:22:13 UTC, Vladimir Panteleev wrote:
>> I would like to ask, what can we improve in our tooling and infrastructure to lessen the burden on release czars?
>
> Merge dmd, druntime, phobos repos.

If I was release Czar, even DMD would just be a dub package. Let them be managed packages that can be selected by version, with many options! That, my friends, is the way.

On Thu, 18 Jun 2015 13:51:05 +0000, Etienne wrote:

> On Thursday, 18 June 2015 at 12:37:15 UTC, Dragos Carp wrote:
>> On Thursday, 18 June 2015 at 02:22:13 UTC, Vladimir Panteleev wrote:
>>> I would like to ask, what can we improve in our tooling and infrastructure to lessen the burden on release czars?
>>
>> Merge dmd, druntime, phobos repos.
> 
> If I was release Czar, even DMD would just be a dub package. Let them be managed packages that can be selected by version, with many options! That, my friends, is the way.

but there is still no package manager, though. dub is not a package manager. and it's impossible to write one without scraping code.dlang.org too, as there is no API to get package list, or packages that was updated since the date.