Re: [Tails-dev] Porting Confidant Mail to tails

Delete this message

Reply to this message
Autor: intrigeri
Data:  
Dla: mike
CC: The Tails public development discussion list
Temat: Re: [Tails-dev] Porting Confidant Mail to tails
Hi,

sajolida wrote (11 Feb 2015 14:35:25 GMT) :
> One of the condition is to be packaged and available in Debian
> first. Is that the case already?


Good question. Let me add something.

>From my perspective, it looks like a new messaging project appears

every two months. Most of the time it is labelled by its authors as
"secure" (which is disputable in itself, but even putting that aside:
not very convincing in the absence of a clear threat model and
security discussion, as we see too often in this area).

Quite a few of these projects are asking us to include their software
in Tails. The thing is, most of the time these projects are not
interoperable with each other, and often not with the existing SMTP
email system either. This, combined to the fact that we cannot include
them all, forces us to pick a winner. IMO it's a bit too early to pick
a winner, the dust hasn't even started to settle, and even the
projects I'm personally trusting more to deliver and maintain
something usable are not ready for prime-time yet.

So, my general recommandation to you, and anyone running a similar
project and willing to make it easily available to Tails users, is:

1. Have your software in Debian, and then in $stable-backports
=> Tails users can easily install your software without relying on
yet another 3rd-party software distribution mechanism.
2. Make it so it works as much out-of-the-box as possible in general,
and particularly on Tails
=> Tails users can use your software almost immediately after
installing it
3. Document exactly what directories and files Tails users should
persist in order to save their configuration, accounts, etc.

Once of all this is done, all interested Tails users can relatively
easily use your software, report bugs to you, and help improve it.
Same for other similar projects. And at some point, I guess we'll have
to decide which one(s) we want to actively support and include by
default. Possibly as experimental features initially.

Cheers,
--
intrigeri