[Freedombox-discuss] Network effect

[Joshua Spodek]

> a) Did little more than host a user's personal mailbox (ordinary
> mailbox - no extra fancy encryption, alternatives to DNS, etc.)
> 
> b) Booted, was solid has hell, was inexpensive, was a solid foundation
> to build on in the future, and had some minor privacy and feature
> advantages over alternatives like 3rd-party hosted email.
> 
> That's a freedombox that does almost nothing at all (other
> than solve some very fundamental problems without which 
> nothing else matters).

Something functional that doesn't break anything later is a great early
target.

FB, like any project, will develop uniquely, but I propose comparing it
with other projects that succeeded or not and using those models to put
FB on track for success.

Regarding usage and adoption, I suggest comparing with Wikipedia. When
Wikipedia began it had few readers, little content, poor quality, etc.
But it distributed under a free license and the people who did use it
(mainly geeks, I believe) were motivated to contribute.

If Wikipedia had tried to include things like having everyone cite
sources, enforcing writing quality, having a more mainstream
distribution of articles, and so on, it would not likely have succeeded
as well. I understand Nupedia would have been better by those measures.
Now Wikipedia has all those things, at least enough for most users, even
though it didn't start with them.

As another comparison, I propose the Linux kernel. I understand the 0.01
release wasn't very functional, but it was free and worked enough (plus
it had the GNU environment to work in). I don't know much about kernels,
but I understand Hurd was intended to be better than the Linux kernel,
but the one that was released with minimal functionality for some people
took off. Today's Linux kernel has enough functionality for huge numbers
of users.

That's the beauty and effectiveness of freedom in a community that likes
to build and fix things.

I don't know the best direction to choose, but I'd like to propose
people writing about the minimum functionality for some sub-community to
use it. Even if it's only useful for a few thousand people (or even a
few hundred, which we already know exists), if those people include a
few people who like to add packages or improve mesh computing now that
they have an operating mesh, that community will turn FB 0.1 into FB
0.5.

I propose that the most important consideration early on, besides
providing minimal functionality for a community that will bring it to
the next stage, is not preventing future essential functionality.
Knowing a huge wish list or FB that spits fire is important to motivate,
but the biggest value of such a list is to know what not to break.

In addition to the FB with everything for the masses list, I suggest
another list: what is just barely enough to be bootstrapped to the next
level -- the equivalent of the 0.01 Linux kernel or Wikipedia's first
interface.

I believe the best way to create a FB with everything in it is to create
a FB that works at all, no matter how ugly or limited in functionality,
as long as the software on it is free, it has enough functionality for a
community that will add a bit more functionality, and it doesn't break
anything later.

Likewise, I believe the best way to create a simple and intuitive
interface is to create any interface at all that works for the community
that uses it in the beginning and to improve it from there. As usual,
you just have to make sure not to prevent anything essential later. I
think the best way to involve great user experience people is to have an
experience at all.

These are just my views on getting started. I'm happy to be pointed out
mistakes, problems with them, or counterexamples where starting with a
low bar hampered progress. But I want to point out their main
consequence, which is a much lower bar for getting started and building
on our overwhelming advantages -- freedom and community. If we
understand how low that bar can be and thereby attain it earlier, we may
get meshes of FreedomBoxes working earlier, growing faster, and the
software growing and being coded and debugged faster -- like the Linux
kernel did relative to Hurd and how Wikipedia did relative to Nupedia.