Re: Kickstarter was not successful... but it did help things...

[Dr. H. Nikolaus Schaller]

Am 20.12.2013 um 12:28 schrieb David Chisnall:

> On 19 Dec 2013, at 21:17, Dr. H. Nikolaus Schaller <hns@goldelico.com> wrote:
> 
>> So if not one person is standing up an saying "go there", we need some other
>> means. E.g. a democratic one. Like an opinion poll and majority votes. Or we
>> do a vote to empower a trustworthy person to define the overall directions 
>> for
>> e.g. one year.
> 
> Leadership requires followers.  Gregory can put on his GNUstep Maintainer hat 
> and say 'we should implement UIKit',

I would suggest that he e.g. says (after discussion, opinion poll etc.): "it is 
very important to work on or finish UIKit next year" and that would give 
guidance to potential contributors who look around that their contributions are 
important. Currently, I think there is no such "direction" and therefore it 
needs people like Riccardo proposed. People who come, look around and 
immediately have an idea where there is a missing piece and then start to work 
on the missing link.

> but it has no effect unless someone actually does the work.  Implementing 
> UIKit is more work than one person can do by themselves.

It is IMHO a hen and egg issue. We always complain that we have no developers, 
but can't tell potential developers where they could (not should or must) put 
their efforts in. GNUstep has a so high complexity for the new-comer that we 
must give some guidance of that type.
> 
> I can give you my perspective on the role of project leadership in the open 
> source world as a member of the FreeBSD Core Team.  FreeBSD elects 9 people 
> (from around 300 active contributors[1]) to nominally run the project.  We 
> can set directions, but we can't actually make anyone go in that direction.  

But from your description I assume that the 9 people do set directions?

> but we can't actually make anyone go in that direction.

Yes, that is common for all community projects. But some are more successful 
and others not.

> The closest we come to being able to do that is by working with the FreeBSD 
> Foundation, which has a budget of around $500K - $1m per year) to fund 
> individual projects that we think are of strategic importance.  Beyond that, 
> the most important thing that we do is talk to companies that are interested 
> in FreeBSD and ensure that they end up talking to the right people to get 
> their jobs done.  The only real power we have is the final say on who is 
> allowed to commit to our repository (and, as XFree86 showed, that power 
> doesn't last very long if you abuse it).  
> 
> GNUstep is a much smaller project.  We had almost all of the active 
> developers at Cambridge over the summer and we fitted in one of the smaller 
> meeting rooms.  Getting a consensus on a good direction is easy.  Getting 
> people with the time and motivation to implement it is much harder.  All of 
> us work on GNUstep because we have some specific need, or as a hobby.  I 
> maintain the runtime and the support in clang because I want a solid 
> framework for experimenting with optimisation and cross-language 
> interoperability research.  I work on Étoilé because I want to eventually 
> have a desktop environment that doesn't suck, but that's further away from 
> things I get paid to do.  
> 
> Various people work on Foundation because they use it in products or internal 
> systems.  Very few people ship products using AppKit and so it tends to be a 
> lot less well supported.  The only thing that the GNUstep leadership can do 
> to improve this is try to find new active contributors (of either code or 
> funding), and this requires finding either very large numbers of users or a 
> smaller number of companies that want to build products or services on top of 
> AppKit (or UIKit).  
> 
> Talk on a mailing list is cheap.  GNUstep is a community that is very open to 
> accepting patches.  There are numerous examples over the last year or two of 
> people getting entirely new projects started in the GNUstep repositories 
> (e.g. CoreBase).  It just needs someone willing to do the work.  If you're 
> volunteering, that's great.  If you're complaining that no one else is, then 
> you're not contributing anything useful.
> 
> Open source projects are not created for users, they're created for 
> contributors. Contributors may be ones who donate code, artwork, 
> documentation, or money.  Hopefully they're also users.  Users are important 
> only in as far as every user is a potential contributor.  If you want to set 
> an agenda for ANY open source project, you need to contribute.  

I think the potential contributors are looking for an agenda set by the 
existing contributors.

> 
> David
> 
> [1] You have to have made one commit in the last year to be eligible to vote.
>