[Opensim-dev] The notion of "core" ... looking ahead

[Morgaine]

Greetings.

Three points of fact:


   1. Opensim is now in Git, a *distributed* SCM that promotes
*distributed*development.
   2. Opensim devs have declared many times that Opensim is not a product,
   but a platform or toolkit from which products can be made.
   3. Opensim distros have started to appear (Diva++), consistent with point
   #2.


These 3 points taken together suggest the following rather likely course of
future history:


   - Git will be used in the manner in which it was intended.  In other
   words, there will be an explosion of Git community repos featuring personal
   branches created by Opensim user/developers outside of the core group, in
   much the way that happened with the LL viewer.  It's likely to happen even
   more strongly in the case of Opensim, because Git promotes this and because
   Opensim code is already nicely modular, which cannot be said of the LL
   viewer.


   - As happened with community viewers, many Git community repos will gain
   high reputations for new features, better performance, more robustness,
   expanded data types, higher scalability, fewer barriers to open teamwork,
   alternative interop models, better APIs, and a hundred other things that an
   extended community can tackle but which the small core team has never
   thought of, or not had the manpower to pursue.


   - Opensim distro builders will build their distros from all the best
   features available in all the best known and most respected Git repos,
   cherry picking to make their distros special in whatever way suits them.
   Distro builders will of course also provide their own Git repositories,
   swelling the repo numbers even further and giving them the prestige of a
   good distro name.  The Opensim equivalents of RedHat and Ubuntu will emerge,
   both as distros and as companies, and will become formidable.


The above doesn't require much vision because it's almost certain to happen,
simply because the tools are right, the incentives exist, people like doing
their own thing, and the precedent offered by the community viewers is very
strong.  The only big uncertainty is to what extent it will happen, and how
much control the core group will retain amid the plethora of distributed
repositories.

The latter is very hard to predict.  However, two extreme cases might give
some idea of how things might pan out:


   - If the core group remains closed, secretive and exclusionary, this
   promotes the emergence of more respected upstream alternative repos as
   replacement Opensim Git masters.  If disputes like the current one get
   really bad, there will be wholesale forks of core, destructive competition,
   politically driven non-sharing, and very damaging press and public
   perception.


   - If the core group becomes open and transparent, and embraces
   distributed community development for core features, this promotes the role
   of the core repo as the single (or at least the leading) upstream master, a
   respected concentrator of the best features from broad Opensim community
   development.


I have a strong predisposition for openness so please take this advice with
a pinch of salt, but I believe it's correct nevertheless.  If the current
core group wishes all the accolades and respect that come from a highly
popular and well run community open source project, I believe that the right
course of action is to become *organizationally* open and transparent as
well.

Perhaps reaching version 1.0 and creating an open foundation might be a good
time for that to happen.

Regards, and much admiration for the great achievements so far. :-)


Morgaine.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20091020/f2357e59/attachment-0001.html>