glob2-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [glob2-devel] Revive Glob2 with a kickstarter


From: Linus Probert
Subject: Re: [glob2-devel] Revive Glob2 with a kickstarter
Date: Sun, 15 Dec 2013 13:09:10 +0100

SDL2.0 is out now and compiles for almost any platform... Wouldn't that be of interest?
Combined with a solid testing platform such as cppunit which I've worked quite a lot with it should build a better foundation then the current one.

//Linus


On Sun, Dec 15, 2013 at 10:20 AM, Stéphane Magnenat <address@hidden> wrote:
Hi Leo,


Thanks for sending this email. It hurts me as well to see these bugs
that do not get fixed, and I would love to see glob2 revived!


For now, I do not have much money or coding time to devote, but I can:

* provide some http hosting if necessary (not https unfortunately)

* provide a "postmorterm" experience about the game development to help
with new design decisions.

* provide expert knowledge on state-of-the-art path find algorithm in
open space including varying costs and all sort of goodies (even on 2D
manifolds on 3D spaces but probably that is not needed).

* provide Python software for co-routine-based state machine [1], this
could help a lot with the units state machines. While that code is in
Python, the idea could be reused (see a Scala example here [2]).

* be available to discuss and test the gameplay.

* provide/implement maybe some learning for AI once the engine is in
place (see [3] for a great example for Starcraft).


I fully agree about targeting Android tablets, but I think that the game
should run on normal computers as well. So technically, I see different
options, I give some pro/cons below:

* Blender/Python. There is a preliminary port of Blender game engine to
Android [4] with at least one working game [5]:
  + very integrated workflow
  + great and accessible language (with numpy available for fast math)
  + also available on normal computer
  - blender game engine does not seem official for Android yet

* Native Java/Scala + game engine:
  + good experience for Android
  + Scala is a nice language
  - might not be a great experience on computers (but depends on engine)
  - Scala might be complex for newcomers

* Qt5/_javascript_
  + works well with both computers and Android
  + great Qt API
  + easy to mix C++ for core algos/_javascript_ for game design
  + I have a CMake-based build chain for that
  + Scripting could be switched to Python if needed [6]
  + Could reuse part of the existing codebase if necessary
  - default _javascript_ maybe not as suited as Python/Scala
  - more things to do by hand


So the engine question is open for discussion, but of course the more
important question is what the game should be. This is of course a
question of art/feeling, so I give you some of mine:

* Keep the same basic game system, but with added dynamics in the world:
  - height map world (hence 3D graphics, or mixed 3D map/bitmap units)
  - free poses for units (not on a grid)
  - dynamic environment, with some rivers, etc. (dams ? :-) )
  - meteo?

* Rework the gameplay to have more variability in games, I would do this
by enhancing the dynamics of the world, for instance by adding
additional "hidden" maps such as pollution, etc.

* Listen to players for their wishes :-)


On the technical part, here are the main changes I see:

* use of an easier language than C++ for game logic, we might want to
retain part of the core code in C++, but that depends on the
engine/infrastructure choice.

* co-routine based unit AI instead of hierarchical state machines, as
the latter was the main reason the engine was so hard to understand,
this is a point I find very important.

* the network part is open, not clear how the current system would work
on mobile devices, maybe for that we should run the whole game on server
and only stream what is currently "being seen"? This brings the question
on whether the game should be browser based actually. This is an open
question, I am not really convinced of that myself.

Here are my thoughts of the moment, feel free to comment :-)

have a nice day,

Stéphane


[1] This was developed based on my glob2 experience:
http://wiki.ros.org/executive_teer
[2] http://jim-mcbeath.blogspot.ch/2010/09/scala-coroutines.html
[3] http://emotion.inrialpes.fr/people/synnaeve/phdthesis/phdthesis.html
[4] http://www.blendernation.com/2012/06/26/blenderplayer-for-android/
[5]
http://libregraphicsworld.org/blog/entry/sokoban-garden-android-game-created-with-blender-inkscape-gimp
[6]
https://ep2013.europython.eu/conference/talks/developing-android-apps-completely-in-python

--
Dr Stéphane Magnenat
http://stephane.magnenat.net

_______________________________________________
glob2-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/glob2-devel


reply via email to

[Prev in Thread] Current Thread [Next in Thread]