[tor-talk] Phones for Tor

[Dash Four]

grarpamp wrote:
>> This phone appears to be Windows-based.
>>     
>
> We have some trust in the MS stack concerning
> ability to execute code and move packets properly.
> Sniffing and sending the cleartext... that's an uknown
> but is reasonably verifiable by watching the network.
>   
You trust M$/Windows if you want to, but I'd steer clear of them 
whenever I can.

>> I see that they are banging on about their FOSS on
>> every web page on their site
>>     
>
> They give away the source. There's some blurbs about verifying
> their published binary hash with what you compile.
Yeah, but it is still a Windows source (besides, there is no mention 
whether they distribute the whole source of their product, including the 
Windows components/API, or just their source code alone) and I certainly 
don't have the resources/manpower to audit that stuff, which is why I 
prefer Linux-based systems - it has always been open source, it has been 
around for more than 15 years and it has a large number of community 
members who have contributed to it and made it for what it is today.

>> that the whole software is based on the Windows-platform.
>>     
>
> I don't like cryptophone due to the cost and non-community
> model. But they do offer an Android unit now.
>   
Care to point me out to a link? There isn't anything on their web site 
that I could see, offering Android-based unit.

>> Interesting feature is the Baseband firewall
>>     
>
> I saw that but didn't get what it is. Please tell...
>   
The term "firewall" is a bit misleading, since this feature is supposed 
to prevent over-the-air attacks by your network operator/malware 
"specialists" (see 
http://www.theregister.co.uk/2013/03/07/baseband_processor_mobile_hack_threat?page=2 
for a brief introduction to this). More advanced versions could also 
filter the modem commands sent by the network operator/others. The sad 
thing is that I have not seen full FOSS implementation of this yet.

>> Don't even know if anyone has truly audited android.
>>     
>
> Unless it involves money or rep, auditing is largely a myth.
>   
Indeed. To me, Android is the lesser of the two evils - ideally, I would 
have liked to have "standard" Linux on my phone (as far as I know, 
Ubuntu are the only ones which could offer this, discarding Jolla for 
the time being as they have not made anything official yet), but this is 
a pipe dream, so Android is the next "acceptable" option - at least for now.

The other two options, as far as I can see, are either having Windows or 
closed-source/propriety system, both of which are a no-no for me.

>> There are some crypto programs you can install but it requires
>> the other party to have the program as well.
>>     
>
> This is not a problem in this community.
> And a proper app would recognize your incoming number
> and use that app when you call people who aren't techs
> (friend/family) but told to install it under threat of no calls.
>
>   
>> I'd have better luck buying burn phones for people than
>> getting them to install software and use it properly..
>>     
>
> For them, yes. For you, no, your graph will instantly point
> to you. With that, encrypted content is your last bastion.
>   
Apart from the "locked-in" system (which, again, appears to be 
Windows-based) and the baseband firewall, I don't see anything they 
could offer, which isn't already present.

Secure communications could be accomplished with existing 
tools/protocols (zrtp/ssl for media, tls for signalling) and there is a 
plethora of Android-based apps already offering that - OSTel/CSipSimple 
to mention just two such apps, so I don't really see any benefit of 
Cryptophone at all.

>> fancy menus which don't tell me much
>>     
>
> As in my former note, all we really want is opensource voice/SMS
> encryption over the cell network, preferably without a data plan
> (but not required).
> Because cell's coverage area is better than wifi (which we can
> already use for crypted wifi to wifi with any old app of the day,
> (provided access to the mic and speaker) but not to interoperate
> with cell, see the former data plan for that).
> Everything after that is likely to be much easier... full disk encryption
> of data, call lists, texts, mails, metadata, etc.
>   
Can't you accomplish this with existing "standard" Linux tools? I could 
think of a number of packages in Linux, which deal with disk encryption 
and are quite good for example.