[liberationtech] a privacy preserving and resilient social network

[Eleanor Saitta]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 2013.06.28 15.20, Yosem Companys wrote:
> I want to commend Alirezza.
> 
> It's very important for us to welcome all those who design and
> conduct research on liberation technologies,  I commend him for
> trying to build something of importance to all of us and dealing
> with the inevitable issues that result from doing so.
> 
> One of the great things of this list is that others with 
> liberationtech experience are kind enough to give feedback to
> folks like Alirezza.
> 
> As always, however, let's try to keep the discussion constructive. 
> Rather than simply point out why certain things may not work,
> outline your ideas about how these limitations could be overcome.
> Everything has a solution, if only we brainstorm the problem long
> enough.
> 
> It's one of the best ways we can all learn from one another and 
> continue to make progress in the field.

While I don't disagree at some level, we get a fairly complicated
architectural proposal coming through here on average of once a week;
more like once every few days lately.  Providing a thorough diagnosis
of how to fix said architecture while maintaining the quite often
incompatible goals of the original proposer would eat all of the time
of many of the people on this list.

There is, in fact, value in saying "here are the things that you have
not thought about that present problems with your design", especially
when they are in fact general issues that face many of us.  Hopefully,
folks who are doing new work in this field can look at the problems
that previous architectures have run into along the way and design
systems that don't have those issues.  This is standard engineering
practice.  If a solution is obvious while I'm writing a note like the
one I wrote, I'm happy to propose it, and do.  Sometimes however, the
answer is "I'm sorry, you're solving the wrong problem."

I think it's especially important to point out problems when systems
are presented as something that might not be purely just research
code.  I'm all for research prototypes of all kinds of crazy
architectures that we might learn something from, as long as we don't
try to deploy them to real users.  If something sounds like they might
be looking at taking on real users at some point (such as when they
start worrying about user content), then I think we have a
responsibility to make sure that errors are pointed out -- which is
why I'm taking the time to make the posts I have.

E.

- -- 
Ideas are my favorite toys.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (MingW32)

iF4EAREIAAYFAlHOavEACgkQQwkE2RkM0wpJiAD+Kf0ign7FmOY7S+EcGnJt5TGo
V6eysUhwHDv7qyFAXW8A/iG0D1cJN8EoL7hBejshKe2UUedm8RVafUJt2rtSaTKS
=zRPO
-----END PGP SIGNATURE-----