[Scummvm-devel] Adjustments to the ScummVM engine development and the project scope

Eugene Sandulenko sev at scummvm.org
Mon Jul 18 21:39:10 CEST 2016


Hi Team,

This is the second letter which I am sending to you after discussion with
wjp.

In the last years, since we moved to the GitHub, we gradually changed our
development model. The major difference was that we raised our bar for the
new engines significantly. In short, only the practically finished engines
were accepted.

As a result, it diverged the development outside of the tree, often leaving
the engine authors one-on-one with a colossal work of implementing a fully
functional engine.

Now we would like to bring it back to the times when we were developing
SCUMM, Simon (AGOS), FOTAQ, Sky, Sword1/2 and other relatively early
engines.

The current rules for the engines to be included are becoming the following:

   1. Any ScummVM developer or person vouched by at least one ScummVM
   developer is welcomed to develop their engines in-tree.
   2. That means, that even the half-baked engines or starting ones are OK
   to be merged or even started right in the tree.
   3. The engine inclusion is announced on the scummvm-devel (if developed
   from the scratch) or passed as a PR with 2 weeks provided for any
   objections or comments.
   4. However, if the engine is not yet complete, and the author is not
   active for 6 months or more, or when the engine is not being developed for
   12 months or more, the engine could be removed from the tree. Most probably
   we will create 'scummvm-attic' repository, which will be synced with the
   main tree right before the removal of such engine.

This hopefully will make the ScummVM development more vivid and pronounced,
the engines will benefit from the buildbot, and the porters could try the
new engines earlier on their platforms.

Which also leads to another topic. The project scope.

ScummVM started as a reimplementation of the SCUMM engine, but by the 0.2.0
it has been including the AGOS engine as well. In the years of development
we gradually faced with the engines which embrace non-adventure genres, and
those games were accepted. By these days we have 56 game engines in-tree
which support the vast majority of the most popular 2d point-and-click
adventures. This lead to the point that most of the ScummVM developers have
their best adventures supported and their interest got diminished. With the
same time, there is a big interest in the developing of games from the
adjacent genre, which is RPG.

Thus, I would like to announce that the scope of ScummVM now embraces both
2d point-and-click adventure engines, as well as 2d RPG engines.

That means that the Dungeon Master GSoC task could be merged right in-tree,
and also the current work of Paul Gilbert aka dreammaster with the Xeen
engine could also be merged. However, the fully 3d games are out of scope
due to the technical reasons and completely different requirements for the
backends and infrastructure.

To sum up:

   - Now the engine development is welcomed to be in-tree as early as the
   engine author wants (with the requirement to be active).
   - The ScummVM scope includes RPGs.

As a side note, the official scope description of the project will be
adjusted once we announce support for our first RPG engine.


Eugene Sandulenko
ScummVM Team Lead
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.scummvm.org/pipermail/scummvm-devel/attachments/20160718/2f3af652/attachment.html>


More information about the Scummvm-devel mailing list