Re: [Stegfs-developing] stegfs

[Sebastian U. <seb...@ur...>]

On Tue, 4 May 2004 20:56:33 +0100
john <jo...@po...> wrote:

> Hello everyone,
> 
> We're not getting on very well with this stegfs thing, are we? The
> project has been set up for a year, and nothing has happened.

It's hard for me to say, but you are right.

> OK, here's my idea. Feel free to tell me I'm stupid if you don't
> agree.
> 
> This project is just too hard. A file system like ext2 is very
> complicated. Our aim is to make something even more complicated, with
> encryption and so on added. The Linux developers have put a huge
> amount of effort into making ext2 work well, and we don't have the
> time or resources to match that. Even if we got something to work, we
> would always be playing catch-up: the ext2 developers keep changing
> their code to improve it, so we would have to keep changing our code
> to match. That's why the original version by Andrew McDonald is out of
> date and can't be used any more.

Short and exact description.
 
> So: I suggest we try something simpler instead. We shouldn't make our
> own file system.
> 
> The loopback drivers lets you make a block device out of a file (or
> disk partition). Suppose we modified it so that the blocks of the
> block device were spread around the file at random, instead of being
> in the correct order. Then you could make a file system on top of the
> loopback device, and the blocks of that file system would be randomly
> distributed around the original file or partition.
> Now, with a bit more work, we could make another file system that used
> some of the other blocks of the original file. And then another one,
> and so on. So we've got lots of different file systems all occupying
> the same underlying file or partition. If everything is encrypted,
> then an attacker won't be able to tell how many file systems there
> are: we could show him the harmless files in some of them, and deny
> the existence of the others.

Sounds good for me.
 
> Does this make sense as an idea? I think Andrew McDonald said
> somewhere that he thought of doing something like this, but changed
> his mind, so maybe there's some problem I haven't thought of.
> 
> Anyway, I think this would be a lot easier that the file system idea.
> It's just a modification of the loopback device driver; all of this
> fits into one source code file, so it's a lot less complicated than
> any file system.

Let's ask the guys with the better programming skills. Andrew and
Andreas, could you please make some comments ? Im really sure that my
personal programming skills are not good enough to judge the technical
criteria of this idea, but it sounds logical to me.

Im going to provide any possible assistance, as far as i can, for
everyone who is going to start coding and yes im doing alle the
donkeywork if someone tells me what to do :-). Testing, writing StegFS
documentation and all that :-)

StegFS ist a great concept and there are a lot of people which are
sending mails to me regarding StegFS and the project future. Andreas and
i started the project on sourceforge because we are absolute sure that
a StegFS version for the actual kernel is something the world is waiting
for...

Andreas and i are willing to give everyone, who is interested, project
leader permission because we dont want to slow down the project
progression. It would be a pleasure for us to have an active project
leader beside us who is able to push the project ...

Our priority is to do the best for the project. If someone is
interested, just give us a wink.

-- 

Mit freundlichen Gruessen / yours sincerely

Sebastian Urbach

--------------------------------------------------------
Wir haben massenhaft Leute, die sich fuer Gleichheit
und soziale Gerechtigkeit einsetzen - aber kaum jemand
fuer Freiheit.
--------------------------------------------------------
Prof. Dr.-Ing. E.h. Hans-Olaf Henkel, ehem. Praesident
und jetziger Vizepraesident des BDI, Praesident der
Wissenschaftsgemeinschaft Gottfried Wilhelm Leibniz