it's alive.
and by it i mean full boost-python and boost-serialization integration with our custom npc class. and by alive i mean compiled and sitting there for the moment. awesome show great job! now one step closer to the meaty part:
getting a separate, custom python script to run in-engine for every (pseudo)intelligent entity allowing real-time multi-process editing of all NPCs individually during live testing over a network, and serializing them when finished, to reload for later reuse.
scripting has, until now, been all but nonexistent in source. in my view, such a thing is unacceptable as it bogs down developers with endless, repetitive compile processes and inevitably prevents small tweaks that could add up to drastic improvements, both in design and gameplay. so if it sounds like i've been working on this a while, think about how much time such an immediate and robust scripting pipeline will save us. Months, even! Not to mention the feasibility of testing small changes will lead us---and you---to a much, /much/ better game at the end of the day.
On the effects side of things, I'm looking into implementing a semi-deferred parallax occlusion shader...this shouldn't be difficult, since I almost feel as though I can just adaptively upsample the directional occlusion biased towards the direction of view to eliminate the dependence on expensive soft shadowing. I'll experiment with that later.
Cheers, and happy modding!it's alive.
and by it i mean full boost-python and boost-serialization integration with our custom npc class. and by alive i mean compiled and sitting there for the moment. awesome show great job! now one step closer to the meaty part:
getting a separate, custom python script to run in-engine for every (pseudo)intelligent entity allowing real-time multi-process editing of all NPCs individually during live testing over a network, and serializing them when finished, to reload for later reuse.
scripting has, until now, been all but nonexistent in source. in my view, such a thing is unacceptable as it bogs down developers with endless, repetitive compile processes and inevitably prevents small tweaks that could add up to drastic improvements, both in design and gameplay. so if it sounds like i've been working on this a while, think about how much time such an immediate and robust scripting pipeline will save us. Months, even! Not to mention the feasibility of testing small changes will lead us---and you---to a much, /much/ better game at the end of the day.
On the effects side of things, I'm looking into implementing a semi-deferred parallax occlusion shader...this shouldn't be difficult, since I almost feel as though I can just adaptively upsample the directional occlusion biased towards the direction of view to eliminate the dependence on expensive soft shadowing. I'll experiment with that later.
Cheers, and happy modding!
↧