hartmans at debian.org
Mon Jun 20 23:07:37 UTC 2011
The tricky part about things with distribution, or migration, or
virtualization built in is matching security models. Tools like that
often work well in a cluster of nodes run by the same administrator.
There are a couple of challenges for FB.
The first is the assumption that there is someone running things who can
make informed trust decisions about code.
While we expect to have an owner who can make trust decisions about
their personal information, that's not really the same as someone who
can make decisions about which code to run.
The second is that trust decisions seem to be all-or-none in a number of
I'm not personally that familiar with erlang.
I think that if there are applications that are useful to FB written in
Erlang we should definitely look at them.
I cannot imagine anything about Erlang that would disqualify an Erlang
application being part of FB.
However, I also suspect FB will not be any one language. We seem to be
taking software from where-ever we can find it and using it as
appropriate. We're not going to have a single software system; so we
will be unlikely to have a single language. Thus I think it makes more
sense to look at the applications rather than languages.
More information about the Freedombox-discuss