Re: [GnomeMeeting-devel-list] Skype vs. GnomeMeeting

[m-redlich t-online de (Matthias Redlich)]

Hi Christian and others,

Am So, 2003-09-21 um 22.35 schrieb Christian Strauf:
> > "Which protocols does Skype use? 
> > Skype uses a proprietary protocol which we have developed. We looked at
> > many available protocols when designing Skype and none were good enough
> > for us. We hope you agree!"
> Very cute.
> Well, this basically shows where to place Skype.

true. I don't like such statements, too. But actually I'm not interested
in the guys who are responsible for Skype. I don't really care if they
are arrogant or not. I'm interested in their new technology.

Here is an excerpt from their FAQ which should be a bit more
interesting:

"Is Skype available for Mac OS, Linux, Palm OS, Pocket PC..?
Currently, Skype is only available for Windows 2000 and XP. Stay tuned
for more information about other platforms."

That's why we should think about implementing it in GnomeMeeting or not.

> Apart from that, I really don't think that P2P in Skype's sense of the
> word is the way to go with GM. I strongly believe that basing the calls
> (session initialization and also the codecs) on a common basis (meaning
> loyalty-free open standards) is the right way to go (goto: H.323, SIP,
> etc.). How else would one be able to insure interoperability? If there'd
> be a good open protocol that allows P2P-VoIP/Vid-Conf in Skype's sense,
> sure, why not include it even if there might be some disadvantages. But
> I'm really not aware of such a protocol, especially none that can
> seriously be called an "open standard". Don't get me wrong. I don't
> think that GM must generally stay the way it is right now. I really
> believe that getting rid of ILS and replacing it e.g. with LDAP would be
> a good idea, for example. And I think that, including SIP and also
> looking into multicast with its established open standards and
> mechanisms is the next logical thing to do.

I really agree with you. It doesn't make sense to implement it when they
don't use any (open) standards. Open standards are very important for
Open Source. (or free software) They should be used whenever possible.
Proprietary solutions shouldn't be supported. (if you can choose)

> But clearly, the Skype people don't aim at providing other developers
> with an open standard (at least that's what I make out of what they say
> in their FAQ). And that's why I think that, while I agree with Matthias
> that it's important to keep an eye on projects like Skype and its users,
> it is for the moment not necessary for GM to swim with the P2P wave that
> Skype's creating for VoIP/Vid-Conf. BTW: what's P2P supposed to mean
> anyhow?! If it means "communication of one peer with another without the
> use of intermediary instances (servers, etc.)", then GM already is P2P.
> The only difference is the session establishment. While you need to
> register to an ILS for the moment to establish a random (not direct)
> session with another user (which was also the case with Napster which
> is, despite this fact, commonly said to be one of the first P2P apps in
> the popular sense of the word), this is the only principal thing that's
> different in Skype. The only things that are worth thinking about are
> the privacy element and the NAT-traversal -- having encrypted streams
> between two GM peers (as it can be done with Skype) would be a nice
> thing to have, and easier NAT traversal would also help (even better:
> kill NAT and use IPv6 -- the best way to go ;-) ). But that's something
> one could try to implement and standardize (or better the other way
> around). So, as nice a hype Skype is right now, I really don't see why
> one should "pollute" the Net with another bandwidth-stealing "P2P" app
> if a construct like GM with P2P-connectivity and a centralized listing
> server does a good job. And concerning the users: I believe that serious
> VoIP/Vid-Conf users wouldn't use a program like Skype. Especially
> companies need to stick with standards that are widely supported
> software- and hardware-wise. Even if Skype promises NAT-traversal, I
> really don't think companies would go for such a solution if there's
> another one that can interoperate with a vast number of products that
> stick to the established standards.

I really agree and don't have much to add. We should observe there
development.

Basically I think that their idea with P2P and IP telephony is very
good. Furthermore it seems to be very easy to use. (works behind NAT
without any problems, for instance) Thus Skype seems to be quite
innovative. 

Since Fasttrack already works with Linux/Unix (mldonkey, xmule ..) I
think there is a change that we might have a solution similar to Skype
for Linux too. That's why I sent a mail to this mailinglist.

> 
> Just my 2-¤-Cents.
> Christian
> 
> P.S.: Sorry for the long mail. I sometimes get a little carried away by
> those "P2P"-discussions...

Not too bad. It included some interesting ideas!

Cheers,
-- 
Matthias Redlich <m-redlich t-online de>