dr at jones.dk
Mon Aug 30 09:15:45 UTC 2010
My name is Jonas Smedegaard. I am a Debian developer since the dawn of
this millenium, working on "Debian Pure Blends", Debian-edu and
(recently) multimedia teams, and (a bit) FSO and emdebian teams. I am
self-educated, self-employed, earning a living doing mostly sysadmin
jobs for small heterogenous networks - mostly schools.
On Mon, Aug 30, 2010 at 12:20:08AM +0200, Michiel de Jong wrote:
>I found you while trying to set up an open-source project that explores
>the combination of PGP and Bittorrent, to do things like email, social
>networking, etcetera over encrypted bittorrents. It was thinking of it
>as a desktop application, but after reading this list, i think it would
>be very compatible with the freedom box effort.
It is exciting to hear how this FreedomBox proposal by Eben Moglen have
inspired and lured forward so many novel ideas.
The above one - and others too - sounds quite interesting. I would very
much encourage you to work on it - and will also offer my help getting
your work into Debian. Would be cool IMO for a FreedomBox to embrave
such project when ready.
I strongly recommend, however, to treat this FreedomBox effort not as a
platform for inventing novel pieces, but a platform for stitching them
together when available as individual pieces in Debian - i.e. a task
similar to that of other existing Debian Pure Blends:
* Find novel ideas not yet in Debian but potentially useful in
FreedomBox, and encourage their packaging for Debian,
* Pick suitable selection(s) of *existing* Debian softare packages,
* Decide suitable configuration(s) of these package selection(s),
* Decide suitable maintainance on these package selection(s),
* Help work on improving the packages to be configurable and
mainainable the way we figure out that FreedomBox need it,
* Decide suitable install routines to bootstrap a FreedomBox
* Help work on improving Debian infrastructures to serve the
needed bootstrap routines.
As I see it, this FreedomBox project may not need to ever write a single
piece of code!
Novel code is developed *upstream* (i.e. as separate projects, typically
independent from Debian) and then is packaged for Debian through normal
means of filing RFP and ITP bugreports. Bug fixing and code
improvements are done through the normal means of filing bugreports
against existing packages. Install quirks and maintainance designs are
done by filing bug reports against existing Debian infrastructure.
Sure, we should discuss and coordinate here what we want and how we want
it, and we should coordinate the bugs filed and the package selections
chosen. And sure, we should throw together test routines to easily try
out our composition(s) before fully integrated with Debian.
But no, I don't think we should _develop_ any code as part of this
project. We do not need it, and it we have plenty of challenges just
working from existing 26.000+ packages already fully integrated in the
larger Debian construct.
>I wanted to note a few things, for what they're worth - feel free to
>shoot them down if I missed the point, or if this is outside the scope
>1 - i think there is a need for search. we agree facebook has too much
>power. but what about youtube? if i post a video on youtube, then it's
>public, searchable, and not in any way private. but I wouldn't be
>surprised if they do their fair bit of "spying for free" as well. my
>point is that to really beat all the existing "platforms" (youtube,
>twitter, ...), you need search (although maybe only at a later stage).
I agree that for world domination we need everything, including search,
but do not agree that FreedomBox "needs" search to exist at all.
Please add search to the list at the wiki. And please either find it
invented somewhere, or invent it yourself.
>2 - multiple users may want to use the same plug server. It could speed
>up adoption in for instance a house share if you can create accounts
>for your 5 or 6 house mates on one single plug server.
I believe this is tied to the middle or upper layers of the stack:
* lower layers: machine entity and resources (e.g. disk space)
* upper layers: user entity/entities, relations, and resource use
As such, your point is very valid, but not a show-stopper IMO: imagine
we find a tool working properly to handle a single user but by design
cannot support multiple users - I would then want a FreedomBox that uses
that, until someone invents a similar but multihomed tool.
So again: add it to the list at the wiki (if not already there), and try
locate candidates fulfilling this envisioned need - preferrably packaged
for Debian already, and if not then please file an RFP or ITP bugreport
and list the bug number at the wiki too.
>3 - maybe (I'm not entirely sure about this one myself) one user may
>want to "trust" several plug servers to store his encrypted files. I
>already trust my friends to hold encrypted backups of my data. So I
>might as well store a second copy of all my stuff on the freedom box at
>my parents' house. Also, maybe I also want to run a "freedom box" on my
>laptop, for lower latency when I´m in an internet cafe. I feel there
>must be a flaw in this reasoning, because it completely decouples the
>applications/usefulness from the actual product/hardware, and I don't
>know whether that's good or bad. At least maybe it's something that
>needs defining. Make of it what you will.
Again, it is great that you raise the issue, but please let us discuss
it not in the mindset of implementing something ourselves but in picking
something already invented. Or put it on a wishlist, to encourage
clever enthousiasts to embrace the idea and realize it.
* Jonas Smedegaard - idealist & Internet-arkitekt
* Tlf.: +45 40843136 Website: http://dr.jones.dk/
[x] quote me freely [ ] ask before reusing [ ] keep private
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 836 bytes
Desc: Digital signature
More information about the Freedombox-discuss