Morten, thanks for posting mortem! I deeply hope that this is the postmortem and the project will start to be healthy again.
Thanks @Morten :) ________________________________ AllJoyn seemed to do a little better in this regard. A few things I saw they did that I feel OCF is lacking: 1. A broadly available product like LIFX for developers to play with. 2. Very easy developer tooling to build devices and clients. Visual Studio’s AllJoyn Studio plugin for building and consuming devices and a device discovery tool for interrogating and understanding devices made it really easy to get going. There was supposed to be an equivalent for OCF and it was announced a little over a year ago, but that project has since died out and never shipped (the reason I was told was mainly because of #1). 3. Pre-compiled bits. Don’t make me compile everything. Just give me a package I can reference so I can get started with building what I’m really interested in. Put them on NuGet, Maven etc where developers normally fetch SDK extensions. 4. Blog posts. Lots of examples for all sorts of platforms. Perhaps a contest on hackster.io? 5. Alljoyn had several Community Ambassadors that did various events and help promote AllJoyn. 6. The AllJoyn leadership group did lots of outreach to the community, retweeted/promoted various community work etc. Now I don’t think AllJoyn got a lot of community contribution directly back to the source, but I honestly don’t think that’ll happen until you get a community that’s passionate about building on top of OCF first. AllJoyn just never really got to that point before it was made obsolete. I really do think step one is getting a broadly available consumer product that’s cheap and useful would be a good starting point. I don’t think I ever met anyone building AllJoyn stuff “for fun” who hadn’t started their baby steps with a LIFX bulb. When OCF started, I instantly dropped my AllJoyn .NET/Mono wrapper work and started over on top of OCF. I got a little bit ways, but to be honest got bored and confused because I didn’t have any real devices to test with and learn from. The services in the samples just didn’t “cut it”, and it got boring, plus I felt I was messing with made-up examples like a remote-controlled elevator (would you get on such a thing?) and nothing “real-world”, not to mention I spent a rather large amount of time figuring out how to build those samples and run them (even when I compiled them they wouldn’t run out of the box, because the right DLL weren’t copied to the output folder, which sort of hits #3 above). Just my $0.02. /Morten From: Gregg Reynolds<mailto:d...@mobileink.com> Sent: Friday, June 1, 2018 1:34 PM To: iotivity-dev<mailto:iotivity-dev@lists.iotivity.org> Subject: [dev] Where are the devs? It seems obvious to me (I could be wrong) that OCF is the best way to go for iot stuff. Yet the iotivity project has conspicuously failed to attract a developer community. It's great that the big dogs are paying their people to work on it, but as far as I can see there are very few independent contributors. Why?
