Re: How robust is SpeakFreely?

2002-12-22 Thread Eugen Leitl
As an user of SpeakFreely (7.2 on Windows, stillcan't get my USB headset 
to work properly with SF 7.3 on Linux) I've got the following three items 
on my wish list. (Hey, I wasn't naughty this year. Honest).

1) built-in PKI support, with fallback to clear. Right now it uses some 
   obscure PGP version, and probably doesn't even ask key servers. In
   practise it's much easer to agree on an IDEA of Blowish key -- but it's
   not an out of band communication, and if you don't switch to the same
   key synchronously one party is going to have her eardrums blasted with
   LOUD digital noise. I think it would be simplest to use SSL, with 
   PGP (7.2 doesn't support GPG apparently) support left in for those 
   parties who need it.

   I must stress that currently using crypto means:

   1) people asking you to do some complicated operations on your end,
  while you're unsure why (you just wanted to talk, why does this
  other party asks me this for? what are his motives?)
   2) using some rather technical lingo (have you ever tried explaining
  what cryptography is to a houswife from the Emirates? And why she
  possibly can get in trouble using it? (She doesn't, I looked up the
  crypto regulations for her country)).
   3) if you comply, you get blasted with LOUD SCARY NOISE

   As you can see, here's some heavy negative conditioning at work here, 
   making the average user associate crypto with pushy geeks asking you to 
   do technical stuff at your end and then get blasted by scary loud noise 
   for your pains. Ugh, not again, thanks.

2) Voice Activation with default threshold set to zero as default. 
   Push-to-talk is annoying as hell, and should be the optional mode, not
   the other way round.

3) A realtime display of current lag time (bar and/or numeric) would be 
   very nice. 
   Lag is unpredictable, and varies over time. Ping/pong protocol at meat
   level is very annoying, especially if one have to instruct some 
   clueless party on the other end first, through a link that doesn't
   work like your average phone.
 
4) Did I say three? Four, FOUR things. Even with current small user 
   community one will frequently get talked by new users debugging their
   setup (see points 2-3 to make it easier), or some teenagers who're out
   to annoy. It would be nice to have a realtime public "phonebook" with
   geographical separations, and ability to block connections from some
   parties.

   This point is currently very unimportant, though.

On Sat, 21 Dec 2002, Thomas Shaddack wrote:

> http://www.speakfreely.org/ is a nice, open-source cross-platfor VoIP
> software. Supports encryption by DES, Blowfish, and IDEA.
> 
> Had anyone knowledgeable ever looked at its code? How secure this
> implementation is? Is better to use Blowfish or IDEA? Where are the
> potential holes there?




How robust is SpeakFreely?

2002-12-21 Thread Thomas Shaddack

http://www.speakfreely.org/ is a nice, open-source cross-platfor VoIP
software. Supports encryption by DES, Blowfish, and IDEA.

Had anyone knowledgeable ever looked at its code? How secure this
implementation is? Is better to use Blowfish or IDEA? Where are the
potential holes there?





Re: How robust is SpeakFreely?

2002-12-21 Thread Adam Shostack
On Sat, Dec 21, 2002 at 07:40:34PM +0100, Thomas Shaddack wrote:
| 
| http://www.speakfreely.org/ is a nice, open-source cross-platfor VoIP
| software. Supports encryption by DES, Blowfish, and IDEA.
| 
| Had anyone knowledgeable ever looked at its code? How secure this
| implementation is? Is better to use Blowfish or IDEA? Where are the
| potential holes there?

Use Blowfish, you avoid worrying about if you have to worry about
patent issues.  There are probably buffer overflows, and other
problems with the code.  But its probably no worse than other VOIP
code, and is clearly more secure than code which doesn't encrypt.

Adam


-- 
"It is seldom that liberty of any kind is lost all at once."
   -Hume