[openbeos] Leaving OpenBeOS : David

  • From: "Anil Joy Varughese" <anil_joy2k@xxxxxxxxxx>
  • To: <openbeos@xxxxxxxxxxxxx>
  • Date: Sat, 21 Sep 2002 12:14:01 +0530

Leaving OpenBeOS
It's now been a fair while since I left the OpenBeOS project and I've
had a lot of people ask me for my reasons. I've stayed quiet on the
subject, but the questions keep coming and people seem to have questions
that they feel I should answer. Here I'm going to try and remove the
doubts that people seem to have and set the record straight once and for

First off I'm not going to rehash in detail the incidents that led up to
my leaving. Nor am I going to point fingers. I'm just going to give the
reasons I left and outline what I'm doing now.

It's useful to bear in mind that I've been worked on several open source
projects and have had a fair bit of experience using CVS and mailing
lists to keep abreast of development on projects. I've also had
experience working with others in different time zones and countries
using these tools. Sadly many of the OBOS developers haven't. OBOS is
their first experience of such large projects. Many are good developers
with excellent coding skills, but they are so used to working on their
own projects that they lack many of the "interpersonal" skills that such
large projects demand of people. They're not alone in this (it's
something that virtually all projects suffer from) but OBOS lacks a core
of people with skills to cope with such wide variations in personality
and attitudes. The result was people changing code in areas that they
had only briefly looked at, overriding the original codes logic or
intent - something that other OSS projects regard very poorly, but that
seemed to be accepted and even encouraged in OBOS! 

Additionally there were incidents where items discussed on admin lists
(in confidence) was openly discussed in IRC channels, totally destroying
the credibility of the so called "leaders" of the project.

The principal of having teams for the different areas seems like a good
one but the borders of these became too rigid, and the people appointed
as "leaders" sometimes took it upon themselves to be overbearing in
their decisions. I always saw them more as discussion forums where the
leaders only role was to give overall shape and be the final casting
vote in case of difficulties, but that wasn't how some saw their
position. I'm not sure that in the fullness of time the team's will
survive as they do at present as they aren't as flexible as they should
be and tend to stifle development rather than encourage it.

Having so few active developers also led to problems with people working
in too many areas. The kernel team was comprised of people who generally
worked on their own projects as well as the kernel, normally at the
determent of their original project. I was guilty of this but as the
networking code had gone as far as it could without better kernel
support felt I was entirely justified in making that decision. I doubt
whether the same could be said for everyone involved.

The kernel itself is an enormous project and the group working on it
lack kernel development experience. Tasks such as rewriting the VM from
scratch are interesting and challenging, but run contrary to the aim of
OBOS to head for a release as fast as possible. I have a feeling that
with more kernel development experience on the team many of the mistakes
they have/will make could have been avoided totally. Sadly kernel
development experience is hard to find and the project hasn't really
attracted anyone with such experience to assist. Additionally the small
group of developers that are working on the kernel lack a sense of
direction. Here the conflicting and mutually exclusive aims of working
towards a goal and allowing people to work on areas they're interested
in has been poorly managed.

Surely these are all just growing pains? I'd like to think so but in the
6 months or so that I was deeply involved in the project I saw no signs
that these issues were being acknowledged or dealt with. Several times I
tried to bring these issues into the discussion and failed or met with
only limited success. The reason that these issues need to be dealt with
is that they will stop developers with more experience joining the
project, something that needs to happen if it's to stay viable and ever
reach it's objectives. After having run into these problems one time too
many I decided that enough was enough.

BeOS was one of the best operating systems I've used and deserves
another chance at life, but unless OBOS sets it's house in order and the
developers on the project clean their act up I can't foresee it ever
reaching a conclusion, however sad that may sound.

Does this mean I've given up on BeOS totally? No. I'm still working on
apache and php for BeOS and am working with some others to try and
release a set of binaries that people can use. There is after all life
for BeOS without OBOS!

Other related posts: