Really interested in what people think on where developing is going (really need a reality check for myself because i start to think i am the only one worried).
I will say that there seem to be quite a few pull requests addressing the first two points, but I’ve also seen periods where activity in other areas was lower than now.
If I had to justify the points you are concerned about, I’d say its a shift to re-target the hobbyists which desire cheaper boards. I’m not sure that is the best thing to do, but it does purport to expand the user base.
The move to ChibiOS opened up a host of possibilities with the freeing up of system resources, so I think its inevitable for such feature and support creep to occur. Not only that, but ChibiOS simplified the process of migrating to new boards via hwdefs, and involvement of the community in ChibiOS deployment helped educate in just how this all works.
I’ll admit - I was really nervous when the big push for ChibiOS started and the issue reports started flowing in, but in my observation we’re well into (and approaching the end of?) the teething pains.
I’d say ultimately it depends on how many new developers the first two points bring in, as opposed to the new issues it creates and developer load they generate. So for the time being it may slow development of other features, but hopefully it will speed up development in the long run.
So it looks like i wasn’t looking at the complete picture. Having moved to chibiOS will possibly bring new devs and it unlocks future possibilities. That would be a great thing.
This is really something i didn’t consider. Hopefully other comments will come.
Yes, maybe development might not be as fast as we would all wish (specially the Dev Team) but short of having the resources to pay full dedication to developers, the wish list will not move that fast.
Each one on this community contributes the best way he/she can. If some commercial ventures that use Ardupilot as the fulcrum of their solutions/offerings, would be willing to support faster development, we would be much better. It is not the case, and the existing commercial partners are already a big help for the project to support itself, and bring to us all a first class software solution.
ChibiOS support all existing flight boards and features : DONE
port ArduPilot to a wide range of F4 and F7 based flight boards, including boards with integrated OSD and boards in small RTF racing copters : DONE
software support for new CAN hardware : WIP
Scripting : WIP
SLAM integration for position estimation DONE
VTOL control improvements with a focus on better tailsitter support : WIP
architecture update to align with Rover/Copter : WIP
Object Avoidance : DONE
Motor Mixers
bring heli mixers up to date : WIP
ESC feedback handling : WIP
Flight Mode Improvements : WIP
Autotune : WIP
Rover : DONE
Documentation : DONE and still continue
That should be more accurate summary! (from my point of view, I have surely miss some points)
But yes, once again point out that we support numerous board… Support done by Tridge… And that what he said BTW!
There are even dozen of developpements done by the dev team or submit by community that aren’t in the roadmap that pass…
About, the mention of different board on recent release that mostly false assumption that they are only for specific board. Most fix are for rc radio (ok, we support a bunch… And IMHO those aren’t safe) and profit on all boards! Not only one…
I am preparing a small blog post explain more on what have been accomplished from a dev perspective! It should clear some statements.
On ChibiOS work there was intend to support more board (and that is quite magically done now), but it also unlock us more performance for further developpements and new capabilities incoming.
But thanks for pointing that. I think that we could make an update on this list and maybe ask the community for the most wanted feature they would like to see this year! (I said would like, no promises :-))
As a side note, the way we come up with the roadmap is at the annual AP developer conference we go around and ask each developer to add items to a “pledge sheet” and then we organise the list and publish it.
This method means we miss items from developers who don’t attend (we might do a better job this year and try and gather up pledges from the wider community). The magic of open source is that we get these gifts sometimes from developers we’ve never heard of… the downside is that it actually makes it hard to plan what’s going to happen.