Group: Skype Replacement

From LibrePlanet
Jump to: navigation, search
(Tox: Forgot to add my personal experience.)
(Tox: Added MiTM notes.)
Line 112: Line 112:
 
* No discussion of telephone network usage.
 
* No discussion of telephone network usage.
 
* No official clients yet.
 
* No official clients yet.
 +
* '''No way to authenticate contacts or detect MitM attacks.'''

Revision as of 13:56, 27 November 2014

Skype Replacement is an item on the FSF's High Priority Projects list.

Protocol, not client

We need to make the discussion about protocols, not clients, (think how people say "send me an email" rather than "send me a Hotmail").

The primary functionality of Skype we aim to replace is real-time:

  • voice calls (to computers)
  • video calls
  • text chat
  • voice calls to plain old telephones (via a SIP gateway or similar)

Anywhere over the internet (or beyond), with as many people per conversation, logging, etc. whatever other features (clarify this section as necessary).

The free protocols currently available to do this are.

Voice Video Text Notes
SIP Yes Yes? Yes sessions?
XMPP/Jingle Yes Yes Yes Full presence, group chat, widely deployed
Mumble Yes No? Yes?
WebRTC Yes Yes No Not a protocol, just media Javascript APIs
Tox Yes Yes Yes In early development, not fully implemented

Central Registry?

Apart from technical problems with free Skype replacements, there is a "network effect" problem preventing free protocols (and clients) becoming popular. This might be solved by explaining the problems of Skype to the people and providing the stories about injustice concerning this software.

Free clients

Lists exist already:

Do we need more info? If so it can go here.

SIP? XMPP/Jingle? Mumble? Tox Active development? Platforms License Notes
Linphone Yes No No No Yes GNU/Linux, Windows, OS X, iOS, Android, Blackberry, WinRT GPLv2+
SFLphone  Yes No No No Yes GNU/Linux GPLv3 IAX2 Support
Jitsi Yes Yes No No Yes Java LGPL
Ekiga Yes No? No No Yes GNU/Linux, Windows XP GPLv2+
Empathy Yes Yes No No Yes GNU/Linux GPL GNOME default
Psi
Yate
Mumble No No Yes No Yes GNU/Linux, Windows, OS X, iOS BSD
QRadioLink No No Yes No Yes GNU/Linux (audio only, video planned) GPL
Pidgin No? Yes No? No Yes GNU/Linux, Windows (no calls) GPL
Twinkle Yes No? No No
Asterisk
sipX
Tox No No No Yes Yes GNU/Linux, Windows GPL3

Personal experiences

Ekiga

  • I have used Ekiga in the past and had good luck with it, but only with other Ekiga users on GNU/Linux -- Mattl 15:47, 14 May 2012 (EDT)
  • Ekiga used to work fine for me on the local network, but failed to connect to another user behind NAT firewall. It does not support UPnP mechanism provided by many modern routers and used by almost every torrent-client and by Skype as well. -- v_2e 05:49, 24 May 2012 (EDT)

Jitsi

  • Successfully used to three-way audio-video chat with a pair of Mac users. H.264 only, but VP8 support soon! -- Mattl 15:47, 14 May 2012 (EDT)
  • VP8 support has landed. --Johns 11:20, 30 May 2013 (EDT)

Empathy

  • Empathy is further improved in Fedora 18/Gnome 3.6. I am able to make voice calls on an XMPP network from Empathy to Empathy and from Empathy to the Google Talk desktop client. -- agajan 23:34, 14 May 2013 (EDT)
  • Google browser plugin interop requires a non-free codec (H.264). Video quality is poor without this codec when calling from Empathy to Empathy.
    Even with the new Xiph codec?

I started to install Empathy on Debian Squeeze, using Synaptic. However, marking Empathy for installation resulted in Synaptic telling me that over 120 other packages were required, starting with brasero, brasero-common, cdrdao, cheese-common ... This is lunacy. Either there is a gross mistake in Empathy's packaging, or it really requires most of Gnome, which rules it out for me (I use xfce).

Tox

I've been using Tox for a while on GNU/Linux and moved a few people to it. I'm now using it as one of my only forms of instant messaging instead of XMPP. Jookia 08:42, 23 November 2014 (EST)

The good:

  • Peer to peer with trustless relay servers if not possible.
  • Uses OTR-level encryption between peers.
  • Group chats and calls.
  • Supports group voice calls.

The bad (but planned):

  • Audio support isn't filtered, echos are a problem and voice is somewhat quiet.
  • Multidevice support for one ID isn't possible yet.
  • No mobile clients.
  • Core not audited.

The ugly:

  • iOS support might need reimplementation due to GPL not allowing restrictions.
  • No discussion of telephone network usage.
  • No official clients yet.
  • No way to authenticate contacts or detect MitM attacks.