Mike, can’t say I absolutely disagree, yet I find myself in this weird situation where I can’t fully agree either!
This project has shown, historically and all the way to now, that there is a very large number of volunteers and contributors. Just take a look at statistics on github with wiki documentation for instance, with 182 contributors to date. Or, for that matter, the emerging ideas, propositions and volunteering in this thread.
It’s true that dev team and community members are mostly of the techie type, and as such not necessarily well prepared, or just as importantly not inclined, to freely perform work more related to administration, marketing, or strategic decision making. Yet both the dev team and general community include people who also have those skills, to varying degrees of course, and are or have been willing to volunteer. To what degree is the question, of course, and hence why I do partially agree with you. I do think, though, that divide and conquer is a successful strategy that has already proven to be successful. Also why, incidently, Ardupilot has a funding committee, a marketing committee, partners meetings, drone conference presence, an annual unconference, etc … ) even if actual performance may sometimes fall short . And there are funds available, if the end justifies the means.
Olivier, You misunderstand me a bit. I am saying that Ardupilot is full of volunteers as well as the core developers. But unlike a corporation with staff to assign and shift around for new projects Ardupilot is not like that. The volunteers come and go working on projects they care about and moving on.
It just annoys me to hear someone say that Michael Oborne should drop what he is doing and do a makeover of Mission Planner for apparently a small group of facebook users. Now that was harsh I admit but I really would say that entry level recreational users are not the most important user base to expend resources upon. We certainly need new users but I also see that as a personal evolution from one of the Naze32 family or from a RTF camera drone where the user wants to do more.
I think it is a problem when no body wants to spend any time learning and studying about new things. Ardupilot’s 700+ parameters are a steep learning curve but with that effort comes a great panorama at the top of the climb. It probably is not worth the climb if you don’t need the advanced features that Ardupilot offers. But if you don’t need those features there are easier flight controllers to use with a more limited set of features.
As you well know resources are limited here and judgments must be made for the best return on invested resources. You also know that I don’t agree with those judgments most of the time either. As I have said previously I think Ardupilot needs a dictator. Having worked in IT for decades, a group of brilliant developers often need a strong leader with a more global vision. Otherwise they often must try and get version 9 out before version 1. Or they just can’t see how difficult it is for someone less expert than they to use their product. That is why corporations spend so much money on focus groups etc. Or they don’t see that the organization should be putting time and effort into the eco-system even tho their primary expertise is flight controllers.
I also don’t think it is bad that Nathan lets it be known what he would like to see as changes. But I do think some folks have the concept that Ardupilot is similar to DJI and will and should assign staff to jump on customers requests. There are no customers here at Ardupilot and there is no “they” just us.
It should go without saying that my opinions are just that.
Bounties are an excellent idea. @lvale Proposals aren’t really relevant to this - they look to be more a mechanism for internal funding requests.
There’s an underlying problem however that needs to be fixed first - PR reviews. There’s a huge backlog of PRs and it takes far too long for them to be reviewed and merged. Any bounty system will likely run up against this wall and quickly frustrate the process.
Maybe the first bounties should allocated to clean the PR list …
The main issue on this list is the number of people reviewing and with merge right … there are too few. That is a part of the good amount of work done on autotest
