[Linux-wildo-devel] Fw: Need a new Linux-Wildo dev?

[Gabriele D. C. <dar...@io...>]

Forward from personal mail.

Begin forwarded message:

Date: Wed, 16 Aug 2006 14:50:27 +0200
From: Gabriele Dini Ciacci <dar...@io...>
To: "Scorpyn" <sc...@us...>
Subject: Re: Need a new Linux-Wildo dev?


Hello!

I am happy you write me, looking for similar projects before starting
your own is always the best approach.

I say you in advance that we would be happy if you join us, we are
always open to get a new member in.

There is currently a totally free position for a gtk coder, cause I
already have so much work that I do not feel like to continue to fill
that position too.

If you do not like the idea of gtk coding we maybe can see for
something other, there is always lot to do.

Let me answer to your points just after them, many of them are points
where there was already lot of discussion in the past and some solution
had already been sorted out.

(continue below)

On Wed, 16 Aug 2006 01:40:34 -0700
"Scorpyn" <sc...@us...> wrote:

> 
> Message body follows:
> 
> Hi!
> 
> Short story very short, I've somewhat decided to make an 
> open source version of teamspeak/ventrilo, and before 
> starting the project I looked around a bit to see if there 
> already was such a project or not and found your Linux-
> Wildo project.
> 
> Anyway, do you think it'd be good if I joined your project, 
> or should I start my own? Keep in mind that the goals don't 
> seem to be exactly the same, and it's been a few years 
> since I've done any coding so I'll have to refresh that a 
> bit.
> 
> To give you an idea of what I want to create, here is the 
> output from my first brainstorming session :
> 
> * Test if it's possible to have different codecs for 
> different users to be able to give better sound quality to 
> the broadband users while still making it possible for 
> modem users to be able to listen 

To allow as many users as possible to use LW the best design choice was
a server oriented design. This means that all users connect to a
central server where the data flow (this is the only possible design in
non peer to peer communication, cause peer to peer distributed design
(gnutella, emule, ecc..) do not guarantee the syncing of the flows and
a low delay (notice that voice speaking is very sensible to delay, more
than what one would expect looking at the numbers).
So in your idea the problem is that it's the server that have to
multiplex various voice streams to various quality, decoding a signal
from a user, re-encoding it to appropriate lower quality/band and
multiplex it. This way you lose lot of quality cause of double decoding
encoding, and the server have a huge overhead of cpu and band usage.
This design has proved to be not only unuseful, but unproductive, 56K
users do not voice chat or use Internet for long time, people that use
a lot Internet and want to chat on it instead of using a simple phone
call on some other mobile normally get an ADSL, and ADSL is already
pretty enought for full voice use.

> * More than 1 person talking at the same time should (if 
> somewhat possible) cause that to be re-coded into 1 stream 
> to save bandwidth 

Since LW is server oriented, it's not VOIP or similar, a Massive number
of users is supposed to be able to talk. Think about MMORPG partyes
that share the same channel during a mission.

> * Encryption (on the voice aswell as the login) 

There is already a short design draft for using ssh both for voice and
login, my best friend is quite a famous expert of cryptography and he is
willing to implement it when the project ready.
 
> * No need for web interface for the server - everything 
> should be possible to manage by settings files, the client 
> etc 

Servers with web interface :DDD eheheh
Nono this is not a project for newbies, if you want to run a server you
must be able to run and manage a server. Certainly the client will
include a "start local server" option needed for fast setup, but i
never ever tought to a web interface, to be true this is the first time
ever i hear someone talking of a web interface to a LW server :)
It's an idea :)

> * Perhaps ability to listen to more than 1 port and 
> redirect to different internal servers (ts can do this 
> already) 

Never used TS, so i have no idea of what this feature is for.
Notice that lw has allready some server to server comunication, so you
shodul be able to migrate from one server to the other in the future.

> * Client and server should both be available in linux and 
> windows versions (linux server and windows client would 
> probably have higher priority in the initial phase) 
priority is as follows:

POSIX client and server (so cygwin and mingw are included)

so this include Linux, MacOSX, Windows, *BSD....
If we find someone that has a Wintendo box he will gladly become the
Wintendo mantainer.

> * tsdisp-like overlay thingy that can be customized for 
> various games (same font, borders, color scheme etc to 
> increase immersion) 

No idea what this is. LW has allready a half-done totally skinnable gtk
client, the design has been done by me and afterwards it has revealed
to be an improvement of the one used by winamp and xmms and a
siggly different approach but same feature of winamp 2 design. It is
based on XML configuration file. You can find and run examples from
LW gtk client dir, there is a doc file that give some guidelines.

If you are talking abotu something invasive in the game, this violate
security policy. You can't use a screen area taken by another
application, apart under certain limits, OSD and similar are ok.
Anyway there is no need of a real overlay, LW is studied to be
integrated inside game client with about 10 lines of code, so it's
up to the game to design another screen where you set the preferency.
What I mean is that LW GUI client is just for games not integrating LW,
the client will probably be included in the game itself, will indeed
see for a release to LGPL maybe.

> * Subchannels in more levels than what TS currently 
> supports 

LW design is: control server, where you log in, multiple data servers
controled by the same control server, you can just change data server
by changing room.

> * Channels that are invisible unless you have the rights to 
> be in them 

Current idea is that all channels are public or with a password, i do
not see why to hide an "admin" channel, anyway this is a detail, not a
feature, cause it's just a matter of who you display or send the list
of channels.

> * Possibility of making ppl in the default channel 
> automatically muted and "blind" (can't see other users or 
> channels) 

LW has half-impelented administation features, with op like irc.

> * Better versatility than ts when it comes to the ability 
> to make groups and set their permissions 

op features are irc inspired

> * Ability to autoselect codec based on who is in the 
> channel and the current traffic etc

on the fly codec switching will be possible but planned for the future.
For now it's rather easy to just create a new channel/room with a
different codec and move to it.

> As you can see, a lot of that isn't really necessary to 
> make it work - my goal would be to make something that 
> isn't just a simple alternative to ventrilo and teamspeak, 
> but it should also be better than both of them combined.

LW never aimed to be a drop in replacement for any of them. As you
have seen many features are for the future, notice that RW did not had
any of those features and until it closed it was by far more used than
TS or Ventrillo. Ease of use, successful easy design often makes the
difference and can even win marketing.

> The list of stuff to be implemented that I've read in the 
> Linux-Wildo info seems to be a lot more realistic though, 
> and definitely a good start.

Indeed, the offical list is a realistic list of things that are not
that far from beeing threre (notice that private testing allready have
sen LW working).

> Anyway, get in touch if you think I might have anything to 
> contribute. If not, I'll probably start up something new 
> from scratch.

I hope you can join us, idling on the irc channel on freenode where we
gather is a good start to get in touch and remember that joining a
project is not as easy as starting a new one, but normally it brings
to results.
 
> /Scorpyn

Best Regards
Gabriele Dini Ciacci

> --
> This message has been sent to you, a registered SourceForge.net user,
> by another site user, through the SourceForge.net site.  This message
> has been delivered to your SourceForge.net mail alias.  You may reply
> to this message using the "Reply" feature of your email client, or
> using the messaging facility of SourceForge.net at:
> https://sourceforge.net/sendmessage.php?touser=1576176
> 


----------- 
http://linux-wildo.sf.net
http://www.diniciacci.org

----------- 
http://linux-wildo.sf.net
http://www.diniciacci.org