2023-05-29, 03:28 AM
This is the type of post I hoped I would never have to write but frustration forces me to write.
I did some troubleshooting and discovered that the RPI Zero W that was in my TinyPilot computer was no good. I can't get anything to work on it, even with a different SD card. After spending considerable time setting up OpenPlotter 3, and after Sean's response above, I figured I would try to uninstall OpenPlotter's version of Pypilot and only communicate with TinyPilot directly in case I had two versions of Pypilot running at the same time. As soon as I completed the uninstall, OP3 died. It would boot to the Rasberry Pi logo then go blank. At that point, I had had it with OpenPlotter. While it seems that it is a good program for those who are power users and programmers, it just requires too much admin work and never seems to work right. Nothing is plug and play, everything needs to be set up individually, which, seems to cause conflicts and the need to fix them. There are so many pages that need to be filled and permissions that need to be set that I have found it to be completely unusable. I have tried using OpenPlotter several times but always with the same result. Yesterday I loaded Rasbian and OpenCPN. I Setup GPSD, added maps, installed Pypilot plugin, and Pypilot fired up first time. I was able to get my motor to move in both directions and I am now a very happy camper.
In the end, all I wanted was to use Pypilot to work with OpenCPN. I am not likely going to have an IoT boat and sit and tinker all day with stuff like that. A boat is already a full time job. I think for people looking to quickly get a working navigation system and autopilot system built and go sailing, OpenPlotter is probably not going to satisfy you. The documentation seems to be where it was two or three years ago but the architecture keeps changing. Again, unless you are a fanatic about building computers and know Linux and programming inside and out, you will probably find it very challenging.
I think that for OpenPlotter to be a success, it needs to get some basic parameters working easily for the average person without advanced computer skills. OpenCPN has been around for a long time and Pypilot for a while too. On their own, they work and have worked. They need to work out of the box with OP before anything else if OP is to be a useful operting system. The documentation needs to be at least to a point where people can get a basic system working. Adding additional functionality, such as engine sensors, AIS, cameras, propane sniffers, and especially interoperability with commercial products should all wait till the basics are working for everyone. OP seems to be important on a certain level and I suppose it should go forward. I am only trying to critique the implementation. I would, however suggest that people think about what OpenPlotter and Pypylot should be in the future.
Where is this all going? Is the goal complete interoperability with commercial products even though the corporations that make the products have absolutely no desire for the same? Or is the idea to build capabilities for people without the means or desire to purchase expensive products to have similar or superior capabilities of commercial alternatives? I would argue that the latter is more important. Building an operating system around Signal K seems to be a losing proposition over the long term. Eventually someone will get sued by some company for copyright or patent infringement in order to crush out the competition and that will be it. Or, striving for interoperability with commercial products may make people just give up entirely on open source because the corporations are implementing what open source develops for free. I think it would be better to focus on building open source products that go their own direction independently and leave the companies in the dust. That would create head to head competition rather than wannabe products.
I did some troubleshooting and discovered that the RPI Zero W that was in my TinyPilot computer was no good. I can't get anything to work on it, even with a different SD card. After spending considerable time setting up OpenPlotter 3, and after Sean's response above, I figured I would try to uninstall OpenPlotter's version of Pypilot and only communicate with TinyPilot directly in case I had two versions of Pypilot running at the same time. As soon as I completed the uninstall, OP3 died. It would boot to the Rasberry Pi logo then go blank. At that point, I had had it with OpenPlotter. While it seems that it is a good program for those who are power users and programmers, it just requires too much admin work and never seems to work right. Nothing is plug and play, everything needs to be set up individually, which, seems to cause conflicts and the need to fix them. There are so many pages that need to be filled and permissions that need to be set that I have found it to be completely unusable. I have tried using OpenPlotter several times but always with the same result. Yesterday I loaded Rasbian and OpenCPN. I Setup GPSD, added maps, installed Pypilot plugin, and Pypilot fired up first time. I was able to get my motor to move in both directions and I am now a very happy camper.
In the end, all I wanted was to use Pypilot to work with OpenCPN. I am not likely going to have an IoT boat and sit and tinker all day with stuff like that. A boat is already a full time job. I think for people looking to quickly get a working navigation system and autopilot system built and go sailing, OpenPlotter is probably not going to satisfy you. The documentation seems to be where it was two or three years ago but the architecture keeps changing. Again, unless you are a fanatic about building computers and know Linux and programming inside and out, you will probably find it very challenging.
I think that for OpenPlotter to be a success, it needs to get some basic parameters working easily for the average person without advanced computer skills. OpenCPN has been around for a long time and Pypilot for a while too. On their own, they work and have worked. They need to work out of the box with OP before anything else if OP is to be a useful operting system. The documentation needs to be at least to a point where people can get a basic system working. Adding additional functionality, such as engine sensors, AIS, cameras, propane sniffers, and especially interoperability with commercial products should all wait till the basics are working for everyone. OP seems to be important on a certain level and I suppose it should go forward. I am only trying to critique the implementation. I would, however suggest that people think about what OpenPlotter and Pypylot should be in the future.
Where is this all going? Is the goal complete interoperability with commercial products even though the corporations that make the products have absolutely no desire for the same? Or is the idea to build capabilities for people without the means or desire to purchase expensive products to have similar or superior capabilities of commercial alternatives? I would argue that the latter is more important. Building an operating system around Signal K seems to be a losing proposition over the long term. Eventually someone will get sued by some company for copyright or patent infringement in order to crush out the competition and that will be it. Or, striving for interoperability with commercial products may make people just give up entirely on open source because the corporations are implementing what open source develops for free. I think it would be better to focus on building open source products that go their own direction independently and leave the companies in the dust. That would create head to head competition rather than wannabe products.