Week 44/2012: Roadmap updates, trimming Snowberry
Last week I fixed the plugin related regressions in the unstable builds and started trimming down Snowberry. DaniJ continued revising and fixing the libdeng1 file system and resource manager.
One of the most fundamental goals on the roadmap is getting rid of Snowberry, our frontend/launcher app, making Doomsday an autonomous entity. This will still require plenty of work, but we've been taking big steps toward this already: the "ring zero" mode for changing games at runtime, resource management improvements, and persistent display configuration inside the engine. Now I've started the process of removing redundant functionality from the launcher: the display resolution setting is no longer offered in Snowberry. You still have the option to force windowed/fullscreen mode using the launcher, but you should use Doomsday's Control Panel for display configuration.
On Mac OS X, another oversight with the frontend has been that one has been unable to switch game plugins at runtime because only the chosen plugin's resources have been made available to the engine. This has now been remedied: even if you've launched with a Doom profile, you can unload it in the console and load, for instance, heretic-ext.
There is another roadmap item that should be mentioned: removal of the "j" prefix in game plugin names. If you've been with us since the early years, you'll remember that the project started with my Hexen port called JHexen. Nowadays it makes less sense to have a unique identity for each game plugin as the operating model is that of an engine with game plugins that can be swapped at any moment.
Note that the old "jDoom" etc. names inside add-on resource folders will continue to be supported as we don't intend to make all existing resource add-ons incompatible.
Finally, the roadmap has undergone some revision. Observing our rate of progress, I pushed the autosave and input related items to 1.9.11. We've also decided to not release a 1.9.9-1 patch as the file system/resource management fixes are still underway and are much too in-depth for a simple patch release. This is mostly because we decided to fix the underlying issues in the architecture rather than just apply a band-aid.
One of the most fundamental goals on the roadmap is getting rid of Snowberry, our frontend/launcher app, making Doomsday an autonomous entity. This will still require plenty of work, but we've been taking big steps toward this already: the "ring zero" mode for changing games at runtime, resource management improvements, and persistent display configuration inside the engine. Now I've started the process of removing redundant functionality from the launcher: the display resolution setting is no longer offered in Snowberry. You still have the option to force windowed/fullscreen mode using the launcher, but you should use Doomsday's Control Panel for display configuration.
On Mac OS X, another oversight with the frontend has been that one has been unable to switch game plugins at runtime because only the chosen plugin's resources have been made available to the engine. This has now been remedied: even if you've launched with a Doom profile, you can unload it in the console and load, for instance, heretic-ext.
There is another roadmap item that should be mentioned: removal of the "j" prefix in game plugin names. If you've been with us since the early years, you'll remember that the project started with my Hexen port called JHexen. Nowadays it makes less sense to have a unique identity for each game plugin as the operating model is that of an engine with game plugins that can be swapped at any moment.
Note that the old "jDoom" etc. names inside add-on resource folders will continue to be supported as we don't intend to make all existing resource add-ons incompatible.
Finally, the roadmap has undergone some revision. Observing our rate of progress, I pushed the autosave and input related items to 1.9.11. We've also decided to not release a 1.9.9-1 patch as the file system/resource management fixes are still underway and are much too in-depth for a simple patch release. This is mostly because we decided to fix the underlying issues in the architecture rather than just apply a band-aid.