Dear Core Developers, Contributors,
and Users,
I am sure there are many users of SER
who now ask the question: What the h... is going and how will this impact SER
and how I use SER? I will try to contribute with my interpretation of all this,
hopefully to the benefit of those less familiar with SER. I ask you to bare
with me, this email is a bit longer than the norm...
In open source
development, everybody is equal and on the surface have the same
motives, but in cases like this, it's useful to know where people are coming
from. I, myself, work for a company that uses SER as a component
in one of its products.
My interests in
SER are twofold:
1. The more features and the better
open source SER gets, the better for my company
2. The more carrier-grade features
(stability, robustness, and scalability), the better
When I first started using SER, I realised that #1 was
held down by a very high threshold for first use + a sluggishness to include
external contributions into the project. I also saw that there was a lack
of some carrier-grade features (like load balancing, monitoring, etc).
So, this caused me to participate in the Getting Started
documentation project, as well as to set up
http://onsip.org to lower the getting started
threshold and facilitate access to information relevant for carrier-grade
setups. And I also volunteered to be the administrator of the new experimental
directory for SER modules.
Iptelorg.com and Voice System have both established
businesses with SER as a core of their business model. Before an open source
project gets the momentum (in terms of development community), a company with
commercial interests in the open source project is crucial to the development of
the project. Once the project gets a life on its own, there will be
conflicting interests between the commercial company and the open source
project. (Fraunhofer and then) Iptelorg.com has been this company for
SER.
Iptelorg.com's business model has included building
other components/extending SER and offer this as a carrier-grade platform.
Of course, they don't want to submit their commercial code to the public CVS,
it's the basis for their revenues. However, when the community gets going,
their commercial components start popping up with alternative contributions
(like TLS). What to do? The reactions can be:
1. Ignore it
2. Submit own commercial code when there is no other choice
3. Pro-actively push commercial code to the open-source project and
continously improve own commercial code
Voice System has a slightly different business model,
but also they have their own commercial components and I believe they focus
more on enterprises. Do they contribute their commercial code to OpenSER?
Of course not, they have exactly the same problems as iptelorg.com, but they
need more features (to enterprises), while iptelorg.com needs stability (to
carriers).
So, my interpretation of the OpenSER announcement:
It's a branch with a different release philosophy, release sooner, bug-fix after
release and bring in as much contributions as possible. Daniel-Constantin Mierla
states "the SER code maintained by us will go further [...]
The cvs was created just to ease the maintainance."
However, the communication, both in the announcement, as
well as on the website is another with a language clearly indicating that this
is *something different from SER at iptel.org*. To the casual user, this
means choosing between two projects and also choosing where to contribute.
(To me, this means that allthough I would love to get new functionality
faster, I cannot possibly use the OpenSER project due to my need for testing and
stability.)
To the iptelorg.com
and Voice System guys: Even though SER originates in your Fraunhofer pet
project, SER is no longer yours. Close to 700 people have registered at
ONsip.org. There is a large and thriving community of SER users! This
gives you, the core developers, a responsibility you didn't have before and you
must act accordingly.
Jiri and Bogdan-Andrei: I plead you to talk
together and resolve this to the benefit of the *community*, not your individual
companies. There is no question that the openser.org website brings additional
value to the beginner user (documentation + functionality) and that Jan
Janak's release philosophy has been a benefit to those of us who rely on
stability. Your indifferences MUST be possible to resolve within the
context of the community. Jiri: Be open to how contributions and releases can be
handled without sacrificing stability. SER needs contributions and Voice
System has a point there! Bogdan-Andrei: Try to figure out how you can
resolve this without splitting the community in two projects. You really need
the other iptel.org contributors in making your fresh releases stable as fast as
possible!
If this only boils down to different release
philosophies and how contributions shall be taken into SER, I have the following
suggestions:
- Let loose the Voice System guys on releasing new SER versions with new
functionality, don't keep functionality in CVS HEAD too long
- However, do as ex. OpenLDAP and state CLEARLY which
version is considered stable, make sure that source packages are available for
all versions
- If absolutely necessary, keep two
sites: iptel.org and openser.org, but please don't communicate that these are
two different projects. If the code base is the same, they are not! Ex.
have one site for stable
releases and one for newer releases.
I will be happy to
mediate in the talks if necessary. After all, I'm Norwegian and conflict
mediation is one of the few things Norwegians are known for...
:-)
Best regards,
Greger
g-)