FOSDEM - Open Source VoIP

[Diana Cionoiu]

Hello Kilian,

Since it takes about 2 years normally to get a new software into distros 
i don't see your point.
Usually everyone that asks for a package is doing that after a while.
And actually it does makes sense to have 100 different endpoints, since 
they perform different tasks. For example YateClient has an embedded 
browser which has been added to be able to support call centers activities.
Everyone is just thinking to his needs, but developers in general have 
to make happy a lot of people. And some designs are not proper for some 
activities. For example you can't add support for quasi associated ss7 
links in Asterisk by design, while by design Yate can handle it. You 
can't use Yate for analogic cards because it doesn't supports them, and 
so on.
Also Ekiga is a lovely software that works great for the mainstream 
user. But there are about 5% of the users that may need something else, 
that can't be solved by design. I can think about some weird combination 
of local IVR + conference.
Yes, i do agree that there is a huge amount of wasted time for different 
projects. But each of them has his own purpose. I know a great jabber 
server which was made because developer wanted to learn java 6 and NIO 
support. That jabber server is in my opinion better than the mainstream 
java sever. Even if a project teach some developer to do things right, 
or allows others to exchange ideas and is a great gain.
Is not very efficient but in this case basically we all do some 
research, and in the research field also 90% of the money are wasted. 
But sometimes a great project show up from nowhere, and it becomes the 
mainstream against all the odds, than we realize that the effort was 
worthed.
After all is the same situation with Linux vs Windows. Only 10% of the 
users maybe have a real gain from using Linux, but that's enough to move 
the others.

Diana Cionoiu


Kilian Krause wrote:
> Diana,
>
> On Sun, Feb 18, 2007 at 08:41:16PM +0200, Diana Cionoiu wrote:
>   
>> It sounds very nice to have just one application for VoIP but it doesn't 
>> really work. Take a look on what happend with Asterisk. 
>>     
> -(snip)-
>   
>> I really doubt that we want to lose the diversity the open source provides.
>>     
>
> I don't want to trade any of this diversity if there is sensible use
> being made of it. It's just that VoIP is a difficult matter and takes a
> real long term approach to code, enhance and maintain it. The grain of
> salt I was trying to add was plain that if we'll have 100 different
> endpoint apps that are like 35% complete that'll get us nowhere. Just as
> relying on the one app that doesn't perform correctly for exotic use I
> need doesn't help either. 
>
> So to get my point straight, I'm all for new upstreams and new products
> if there is real need and if the maintainer will know what he's actively
> choosing. Just don't try to talk people into wildly starting new voip
> applications and pumping them onto ftps when they'll just find out three
> weeks after they're not up to the task and when there is still more
> potential in existing upstream groups to participate and enhance these
> from 80 to 90% completion or so. Nothing more, nothing less.
>
>