Request for Comments: Agenda Items for 2020 Summit

The FreedomBox Foundation is pleased to announce that we will hold our third consecutive FreedomBox Summit on Friday, December 4th, 2020. This year, the meeting is fully virtual with participants meeting over video conference. Invited members of the FreedomBox Core Team will join the FreedomBox Foundation’s staff for several hours of discussion about our progress since last year’s Summit and our plans for the year ahead.

The full announcement about the 2020 Summit is published on the website of the FreedomBox Foundation.

Though we have an idea of what we want to discuss during Session I, we want community members to have their voices heard. Similar to last year’s format, we plan on assembling an agenda with your suggestions.

In short, our agenda should reflect the most important issues facing the FreedomBox project as we continue to improve our UI, attract new users, and expand our hardware sales.

  • What do you think are the most important challenges facing the FreedomBox project?
  • What should our development priorities be for the upcoming year?
  • We want FreedomBox to become a mainstream consumer privacy solution. What do you think we need to do to get to this point?

The Summit is one of the best opportunities to bring your ideas for FreedomBox into the limelight. Please post your ideas for agenda items for the 2020 Summit below and also make sure to join the Q&A session on 4th December.

As always, thank you for being committed members of the FreedomBox community!


Also see the email thread on the mailing list.

3 Likes

I would suggest the following topics:

Core development

We have continued to add “Components” that can be instantiated by any
app, and which handle things like firewall ports, user groups, web
server configs, domain names, and certificates. The next step is to
move toward declarative apps, which will remove most of the
boilerplate or template code. Then we can split up the FreedomBox
repository and packaging, so that each app can be developed separately
from the core code base. Finally we can set up an “App Center” to
handle installing packaged apps from the current distribution.

Reachability

We have started collecting some basic information from the user about
their network and router. Next we should use this information to
suggest next steps that user can take to make their apps reachable
from the public internet. We already have UI designs and tasks ready
for implementation for this topic.

Video conferencing

As others have mentioned, we should find an appropriate alternative to
Zoom. Better integration of the Coturn TURN/STUN server may also help
here.

Fediverse

From what I can see, Epicyon remains the most viable candidate
fediverse server to be packaged in Debian. The drawbacks are that it
has few users and few upstream developers. Packaging it will require
some work to be contributed upstream.

Email server

This is a basic infrastructure component that will enable other
features such as password reset and email notifications. I think our
initial implementation does not need to handle every situation. E.g. I
would still find it useful even if I cannot send to Gmail users.

Nextcloud

Packaging Nextcloud into Debian Fasttrack will require 2 things:
initial work to package all the dependencies, and ongoing maintenance
to keep it up to date with latest upstream (at least major
releases). I see it as only an issue of prioritizing where to spend
our time.

3 Likes

Here is the set of topics I can think we can discuss:

  • Partnerships: Currently, FreedomBox not fully utilizing the potential of working with other projects. Hardware vendors’ push will increase our user base while providing a compelling use case for selling their hardware. Software projects should start identifying FreedomBox as an easy way to deploy self-hosted projects. We must work together with hardware platforms and software projects we support.
  • Audio/video conferencing: Both conferencing and distance learning has to be a priority given the pandemic situation and how the world will change after the pandemic.
  • Reachability: This remains the biggest source of frustration and usability hurdle for new users.
  • Debian Pure Blends Work: Potential for getting FreedomBox listed in the Debian installer.
  • Security: I believe we made a lot of progress already but we should continue with our improvements and do better container based security. We can also factor out a separate daemon for privileged actions.
  • Components: Finish the components work, make apps more declarative. Build better ways for developers to create more apps.
  • Fediverse: Add new apps.
  • Nextcloud: This remains an important app for FreedomBox.
  • Email server: We can start work on this based on the existing plan.
  • WordPress: This another application that will satisfy a lot of use cases and improve upon existing deficits.
  • Continued UI/UX improvement: There is still a lot of potential for improvement in this area. We can work on items such as queuing app installs, reducing first boot steps with notifications and pushing event updates from server side instead of page refreshes.
1 Like

Requesting comments and ideas is one thing.

But the consideration and appreciation of ideas may be just as important, if not even more important. Meaning sharing existing and own ideas, leading to questions and answers, discuss new aspects, differences, improvents, etc. …

I have tried to discuss an idea. Some limited but crucial adjustments in the fundamental configuration of the new NAS feature (employing and building on filesystem standards).

My idea is for Freedombox to not just provide dumb network storage space, but a great solution for intuitive collaboration among individual freedombox users and groups (and this with proper, user-understandable, directory-based access permissions that allow working on files together without any permission hassles).

And it further allows for many additional benefits, by e.g.

  • Having collaboration features simply by embedding a filemanager web-application.
  • The users gaining access to all their user data they have in their installed Freedombox apps (if these store files in user/group dirs)
  • Possibly mounting the freedombox user dirs as /home on clients. Maybe even using freedombox as LDAP login server and (temporarily) booting clients from a PXE netboot server app.

But with so few responses I’m really asking myself why and what for.

Here are some broad themes I can think of that are worth discussing during the Summit.

Reachability

Though providing FreedomBox as a consumer electronics device has significantly brought down barriers to adoption, setting it up on the home network is still the most significant challenge faced by new users. We should carefully consider new user feedback and use it to guide our solution for improving reachability.

Sustainability

I want to discuss our current sources of funding for the project and discuss Plan B.
Also want to discuss about the bus factor especially for the roles being played by core developers.

The second edition of our HSK

The number of apps supported by FreedomBox is growing. We cannot limit ourselves to basic apps when more feature-rich and popular alternatives are available. This requires us to upgrade the hardware that FreedomBox HSKs run on. We made good progress in 2020 supporting new ARM64 SBCs that are significantly more powerful than our Pioneer HSK. We should start planning for the next edition of FreedomBox home server appliance.

Growing our community

FOSS projects sometimes cannot handle their own success. As the number of contributors grows, the existing core developers will need to shift their focus from development and do things that have more long-term value.

In the past, I mentioned finding people to play specialized roles in the project. Though I don’t see the need for a 1-to-1 mapping between persons and roles, we should still focus on filling all the roles.

Initiatives like “Contributor Invites” need to be restarted. Barriers to entry for contribution should be identified and lowered.

Big apps

Having popular apps is a big driver for adoption of FreedomBox. It seems like the community is currently asking for three major apps:

  • Video-conferencing solutions
  • Office suite + Drive (NextCloud, CryptPad)
  • Email server

Given my time constraints and multiple responsibilities, I don’t see myself being able to fully develop even one of the above in the next year. We need more people developing apps for the FreedomBox platform.

App maintenance

I have recently posted about some apps that are low-hanging fruits. I mentioned in a previous Summit that the current set of 3 core developers cannot possibly maintain 50 apps. We did significant work on the app development framework in 2020 but didn’t add many apps. This is a good thing. Having dedicated app maintainers frees up core developers to focus on framework development and non-functional requirements like quality and security.

Marketing

FreedomBox’s visibility has suffered in 2020. I hope we can do better in 2021. Lack of marketing is one of the biggest problems for free software in general and FreedomBox is no exception to the rule. This is a horizontal effort that can put a significant dent in many other problems.

UI/UX improvements

Successful free software projects like Mastodon owe a lot of their success to the contributions of artists and UI developers. FreedomBox basically has two selling points - privacy and user-friendliness. We need to invest just as much in the latter.

Debian Pure Blend status

FreedomBox being a Debian Pure Blend is an incidental technical decision and not a project goal (as clarified by Eben Moglen in last year’s Summit). We should do a pros and cons evaluation of being a Debian Pure Blend to see if it’s still worth it.

We have to ask ourselves some difficult questions like whether the enormous (and sometimes Sisyphean) effort of packaging an application like NextCloud or Pleroma for Debian is really adding value to FreedomBox users.

Smaller wins

  • Creating a website on FreedomBox is still beyond the capability of most non-technical users, mostly because of Linux filesystem permissions. Creating an easy process to upload website files or adding an app like WordPress can solve this problem.
  • We collect no telemetry of any sort from FreedomBox devices. In addition to the annual Summit, we should find a way of collecting and acting upon the feedback being provided voluntarily by the community throughout the year. Maybe we would have paid more attention to video conferencing on FreedomBox if we did this. :thinking:

It would also be of large help here, to have a collaboration-enabled /home tree (user-private-group scheme and group dirs), that is used as the basis for all freedombox apps and for the network filesystem access. Even now with just SSH, before a samba /home share or a web filebrowser is there.

Example: App data in /home/<user or groupname>/<appname>, and /home/group/admin/<appname>, or a dedicated app group ( /home/group/<appname>)

IIRC @homer77 has reported a lot and contributed his configurations to make XMPP video chats work.

Maybe that is a sign (to pay attention, and include in development) if there are (advanced) users that care to find a solution, and actually pursue as far as they can. Advanced users may be able to do sysadmin work and configure things, but hacking only to some degree. So they can’t actually implement the complete feature, but still help the devs with experimenting and testing.

Good direct feedback is what’s valuable, telemetry is dismissive and actually deserves the same.

1 Like
  • Video-conferencing solutions
  • Office suite + Drive (NextCloud, CryptPad)

I just like to hint on:
Nextcloud solves both Video-conferencing (with Talk) and Office suite (with Collabora) if configured correctly. So Nextcloud alone would already offer:

  • Cloud Storage,
  • Office Suite,
  • Video-Conferencing,
  • Calendar,
  • Task App,
  • Doodle-like Service,
  • Polls,
  • Notes

and a lot more. Including an LDAP-App which should help to connect Freedombox-User-management smoothely.

So actually Nextcloud would kill a lot of birds with one stone.

I don’t mean it bad, just please consider, though nextcloud does not completely “kill” a lot of low-power single board computers, I found it’s performance is not really satisfactory (pleasant to use) without reasonably fast hardware.

The “nextcloud-pi” distribution already follows that all-in-one nextcloud goal, a drawback however compared to a debian/stable/freedombox seems to be its level of “appliance” stability/updates over time.

Freedombox has the advantage to be able to combine the best and most efficient tools and solutions for each task with a simplified adminstration (in debian).

As far as I know, Nextcloud’s Office Suite feature is in the form of integration with a separately hosted Collabora Office, which needs significantly more resources on the server as compared to CryptPad.

FreedomBox already has a calendar and tasks app, Radicale.

CryptPad can do the following:

  • Encrypted Cloud Storage
  • Office Suite
  • Kanban
  • Whiteboard
  • Polls (in development)

It currently list 8 apps at https://cryptpad.fr

A more detailed post about CryptPad.

Just to clarify the requirements, a web-based office suite is not really the need. A combination of LibreOffice + Syncthing can be used to edit your documents from multiple computers. The need we should address is real-time collaborative editing. I haven’t tried collaborative editing on Collabora Online myself, so I can’t compare it with CryptPad in this aspect.

1 Like

Right, and the best is to allow to easily combine the different and best tools for particular tasks, i.e. different users to be able to work with different apps on a shared filesystem without the usual copy/move/permission etc. hassles. So that every app that supports a file type can be used, while all user access permissions are intuitive and maintained completely automatic.

The desirable practice is simply to just have to go into a shared directory like /home/<username>, /home/group/<groupname>, or (if some app really only supports a fixed directory location) /home/group/<appname>.

Hi folks - I wanted to express an idea that I’ve been working on and I plan to participate in the community more as time and opportunities progress.

I really want to utilize my Freedombox to host a music repository for the public to browse/stream through. Looking on the Debian/freedombox wiki, it seems like an option called Funkwhale is the most prudent solution. However, I am running into hell trying to get it to work correctly. Is there a reasonable way to implement something like this? I haven’t planned to try miniDLNA as I understand that its not accessible from simply visiting my public page, and instead would only be available to internal users.

I’m highly optimistic about the potential uses of my Freedombox Server, and I know it’s going to be a challenge all the way forward into a better distributed internet future, but we do have the tools and capabilities to make it happen.

-Sir.Kit

2 Likes

Agenda proposal:

A central issue for FreedomBox to become a consumer-usable privacy solution is, to overcome the completely fragmented use of the (rightly isolated) freedombox apps.

  • Even were group permission allow it (most prominently for the admin, but also for group specific files), there is no central and easy way to properly access the user files that are created or managed with all the different individual apps, not with some app nor over a network share.
  • Where the apps would actually even support write access, e.g. to upload or edit a wiki page over SSH or a web-filemanager, this actually messes up the file permissions and “breaks” the app.
  • It’s not possible to use one app with or on the files of another app where it makes sense. For example, to sync ebooks, or Transmission files with a local client over syncthing. Again the permissions won’t match the requirements and things get broken.

Examples of user-data files:
Bepasty uploads, Calibre ebooks, photo collections, file attachments sent through XMPP or Matrix, website or wiki downloads, syncthing or infininoted files, etc.

As long as these files are not directly editable or move/copyable, the users will need resort to copying and handling files over external devices or services.

Thus there are multiple reasons to make the user-data easily accessible (without creating problems) by the users themselves directly (web or network share login) and through certain apps where it makes sense.

What do you think we need to do to get to this point?

Introducing user collaboration for NAS and FBX “cloud” apps on central user-data (& simpler backup/external disk mapping).

1 Like

The original thread that made me look into the best solution for a shared user-data filesystem:

It may be easier if you can run it in a separate yunohost instance.

Yet here are other options listed:

I really want to utilize my Freedombox to host a music repository for the public to browse/stream through

+1 with @Sir.Kit as music streaming is one of my top use of self-hosting.

Along with my FreedomBox, I also have a NAS where I installed Airsonic on. It’s been working great, and this is the piece of software that I would love to see one day being made available in the FreedomBox ecosystem.

As a side note, to connect with Airsonic and stream music, I use Ultrasonic, which is also a great app.

Thanks for opening this thread to take the community input into consideration :pray:

Hi all. At the risk of repeating what has already been mentioned, the items at the top of my FBXmas List are as follows…

Wordpress/Static Site Builder
User File Manager
Fediverse Instance Hosting (Mastodon/Pleroma)
Next Cloud
Video Conferencing
Up To Date & Easy To Understand Documentation/Instructions
Online Tutorials via YouTube/PeerTube
Discuss/Forum hosting

Thanks for all your hard work everyone!

CM

2 Likes