Thanks for clarification about the LemonPi/RoseapplePi naming .
I understand that it generally makes sense to enable support especially for devices, where manufacturer does not ship a good/current image already. This is basically where ARMbian does a great job (assumption, I never used an ARMbian image, actually). They support a large number of as well rare devices and even for devices with better manufacturer caring, they bring images with up-to-date kernel etc.
But this is not what we (DietPi) do. We do no kernel/firmware/bootloader coding/development, thus cannot enable support for devices, where not already either the manufacturer or another 3rd party provides a working Debian-based image.
Instead we do software (shell) coding with the aim to enable easy (by times rare) software installations where e.g. Debian repo does not provide a package and/or pre-configuration needs much manual effort, and the same for device configuration.
On top, jep we try to be “better” then other images/distros in terms of image size, idle RAM and CPU usage. So providing an as much as possible trimmed down image/system, but then enable as easy as possible configuration and install of extensions/software, as required.
And jep we have comparisons:
Where did you get the 4000 RPi installs from?
I read 36500 installs, ~8000+ RPi installs, when applying the share between users that provide info (opted in survey), then ~20,000 RPi users: https://dietpi.com/survey/
- Since we cannot track users stopping to use the instance or when doing a fresh reflash, some might not be active.
- On the other hand, these are only v6.9 and above installs and there are still pre-v6.9, even pre-v6.0 users, although hard to guess how much .
With always limited amount of coders/devs/time one needs to focus on the wider used devices over the less used devices. If in theory users would use the manufacturer images, when well maintained and only use 3rd party images, when non (current) from maintainer available, then the install counts for RPi in our and in e.g. surely also ARMbian case should be lower compared to all the OPi/BPi/NanoPi/Odroids installs. But this is not the case. So there ARE other reasons for users to use 3rd party images, some of them I mentioned above, when it’s about DietPi. So even that rare-device users are forced to use more current or even working 3rd party images, they are still so much fewer, compared to the well-known widely-spread RPi/Odroids users.
Yeah basically shit when manufacturers throw many devices on the marked but do not take care that at least one well maintained image is available. I saw official BPi image as a patchwork from an OPi image, leaving many obsolete files and confusing configurations in place. They not even produced a single image, build up from scratch, leaving all this work to 3rd parties like ARMbian. If you thinking about buying such a device, then see those ancient, most likely broken, at least patchworked images, provided via external download servers (google, mega and such, this further shows how “much” value is given to those images), I am not surprised that marked penetration is lower than it could be. Not everyone directly has all the 3rd party images in mind.
If you provide an SBC, you NEED to provide proper firmware and software, either do it yourself or make deals with external kernel developers/image providers that you can then provide on your website. Everything else it ugly short term money making until users realize that they are lost with obsolete broken work…