Including new valuable apps into freedombox

[spotter]

Summary
How to develop a process to incorporate new applications into the freedombox application suite and ensure their sustainability and permanence in freedombox?

Problem
Hi,
I’m a freelance volunteer, and I translated much of Freedombox into Czech language.

In addition, I am working on my idea-project, shortly presented here https://spotter.ngo, which was created since 2016 independently of Freedombox.

In short: I test and localize various open-source software for internet-independent use. For this I used to develop my own backend based on Alpine and my own Python code. All in singe virtual machine file. Further progress became complicated and I decided to adapt to a different type of backend. Freedombox looks like a much better and long term development environment.

I would like to find volunteer system integrators who can help build specific applications into installation packages and test them for use within Freedombox.

These are applications for heavy use by non-profit organizations in crisis situations, first aid, or development assistance. I talk about range apps with GIS, functionality.

I know it can always be argued that every application has prepared its own Docker, so why bother with integration? I have facilitated experience with these server-apps is several years of testing and expensive programming and integration work. Not easy at all. That’s why I argue that you can’t always rely on Docker, or source code in an untested version, to create a immediately user-friendly deployment. In addition: independent on problematic commercial technologies or services.

Is the freedombox project amenable to creating such “specific” sets of applications that would be accessible from its interface?

What approach should I take?

Unfortunately, I am not an active programmer or integrator, so I need active help to create and test installation scripts. Final solutions will be open-source.

Is it better to build everything on fork of freedombox or to integrate it into the main project?

Thank you very much for your thoughts and reply

[njoseph]

Hi @spotter. Welcome to the FreedomBox Forum.

I went through your website and I think I got a high-level understanding of what you’re trying to do.

I agree that FreedomBox would be a better platform for hosting server applications than creating your own custom solution.

All current FreedomBox applications are maintained as Debian packages. This gives us good defaults for security and privacy, which could be missed if installing directly from sources. The effort of maintaining the packages is shared among the larger Debian community. Debian packages can also have some additional security patches which the upstream source hasn’t integrated yet. FreedomBox’s automatic software updates keep maintenance minimal for end users.

I was a volunteer in a team that installed FreedomBoxes in remote Indian villages which had no Internet connectivity. We had to install some applications manually since the corresponding Debian packages were not available yet. We documented the process in a WikiBook called FreedomBox for Communities. To scale up this process, we created custom disk images of FreedomBox with all the required apps pre-installed. Since then, some of those applications were packaged for Debian and FreedomBox.

At least some of the volunteers should have system administration skills for this model to work. They can guide the others. Though the installation is manual, you get the benefits of a stable and secure operating system. With containers, you will be forced to upgrade if your container has a security vulnerability in an OS dependency such as glibc. For example, if you are installing DHIS2, Debian’s security updates will take care of the stack all the way up to the JVM. You only need to worry about security vulnerabilities in the DHIS2 application and its Java libraries.

I don’t think a fork of FreedomBox is required. You can display your manually installed applications on both the web interface and Android app by adding the required data on the FreedomBox server. See “Custom Shortcuts” on this wiki page: FreedomBox/Customization. Maintaining a fork would be too much work. It is better to add any required customization capabilities to FreedomBox itself.

Hope this helps. Thanks for the Czech translations. I hope you find volunteers with a diverse set of skills to make this happen.

[spotter]

Thank you for understanding the intent and the links to help find solutions to the topic.

I think a specific specification can be found for building a standalone map server environment on Freedombox, similar to what is done for offline maps on mobile devices. This would further open up new horizons to leverage other map service dependent applications that have unnecessarily large amounts of data from the internet.

These solutions exist for closed use, which gives me the information that this is a valuable concept that brings the advantage of independence from the services of commercial companies such as Google maps…