[tor-talk] [Tails-dev] Tails Mac support

Maxim Kammerer mk at dee.su
Mon Feb 25 20:10:49 UTC 2013


On Mon, Feb 25, 2013 at 3:33 PM, intrigeri <intrigeri at boum.org> wrote:
>> not only are you completely dependent on an upstream distro's
>> features implementation cycle
>
> I've no idea what misconceptions about Tails and Debian make you think
> this, but this is incorrect in practice.

I really don't know why the moment I mention something about Debian,
people get very defensive and assume I don't know something. Debian is
nothing special, it's just a binary distro that requires no
understanding to use — the reason it's a base for so many forks,
including Ubuntu. The only reason people get defensive is pure
rationalization due to being vested in Debian — like with that
certificates fiasco, which, were it to happen to any less popular
project or company, would result in its ridicule and eventual death
(e.g., DigiNotar).

Now, of course what I wrote is correct. For instance, you don't, and
won't have UEFI support until Debian community decides to implement it
in a way they see fit. Moreover, due to not gathering experience while
working on said support, you will have nothing to do with the
solution, once it's ready. Same with so many other things. You accept
bug reports and maintain related bug / todo pages, knowing full well
in advance that this leads nowhere. This also misleads users, e.g., I
have seen someone on Twitter mention your earlier message about UEFI,
applauding the apparent progress, while in reality you are just
rehashing the same old information.

> First, I fail to see what compiling binaries yourself buys you, in
> terms of your level of dependency on "upstream distro's features
> implementation cycle".

Gentoo is not about “compiling binaries yourself”. Gentoo is a
source-based highly flexible meta-distribution, each component of
which can be easily changed and adapted to specific needs. Gentoo is
as close as one gets to LFS, without having to actually do everything
manually and while keeping decent package management. You wouldn't
understand the advantages just from the description, because in boring
distributions like Debian the developer is still a “user” — you need
to go out of your way to modify system behavior. Debian does not
encourage understanding and experimenting. E.g., I remember you, or
one of the other Tails people asking on IRC: what good is ASLR?
Indeed, how would you know, if the distro you use discourages users
from deviating from stock kernels, to the point where you would
initiate a long bureaucratic process for changing a single trivial
kernel setting that is needed for Tails?

> [contentless propaganda skipped]

> So, yes, e.g. having UEFI support added to Debian Live makes sense to
> me, as opposed to implementing in a Tails-only way and maintaining it
> forever. Sure, it sometimes means we get the feature a bit later
> (which is not that clear in this specific case).

You have no idea what you are talking about. Whenever you *do stuff*,
instead of waiting for someone else to do it, while engaging in
useless “community relationships”, that someone will usually end up
actually using the results of your labor. With UEFI, it's just too
funny — you will likely end up using sbsigntool directly or indirectly
(which is already used in Ubuntu), which contains my patches, which I
added because I needed sbsigntool working properly in Liberté. Oh, and
I learned something new, which was great. But I guess that waiting for
stuff to happen is just as exciting.

> If you intend to go on writing such bold public statements about
> Tails, then I'd rather give you first-hand information that you can
> base your affirmations on. Here we go, then. My experience absolutely
> does not match your assumption. I'm personally quite happy with how
> I've been learning new things when implementing features for Tails, be
> it when writing Tails-specific software, or when implementing the
> feature upstream, or when working to make Debian an awesome platform
> to build the next Tails generation upon.

Without being able to quantify that statement, it is just another
politically correct contentless propaganda. I have read some Tails
monthly reports, and items they are composed of are of the kind that I
usually don't even copy from git commit messages to the changelog file
in Liberté. I mean, it's nice to see people touting Tails features
that were copied from Liberté (like unsafe browser [1], or memory
erasure on boot media removal, or clock setting / whatever else), but
from my point of view, you don't do anything interesting (so I nearly
stopped following your project), which is a shame, hence my replies
ITT. Think about what benefits your project. If you do hard stuff, you
attract other people who can do hard stuff, whereas otherwise you
attract people who know how to tweak settings / apply patches, and not
much else.

[1] https://lists.torproject.org/pipermail/tor-talk/2012-July/024964.html

-- 
Maxim Kammerer
Liberté Linux: http://dee.su/liberte


More information about the tor-talk mailing list