Middle East map, facts, and flags. Physical map of Iraq, map of Turkey, Dubai map. Free map of Middle East for teachers and students. Great geography information for students.

Open and view AVIF images online without installing software. View multiple AVIF files at once, see dimensions and file size, and click to zoom. Works instantly in Chrome, Firefox, and Safari.

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

Just theorizing at the moment, this would clearly be a jme 4+ thing. but slang does support generating all permutations based on interfaces. We could generate them at compile time, and then work with plain glsl/spirv during execution. it of course needs a fair bit of magic, and the big problem is the currently used int defines. i do not know about the feature set of gles / webimage enough to know if we can skip them. in modern opengl if we move to UBO/SSBO this issue is gone. (of course that requires the vulcan implementation first to support osx) The core issue with the current material system is that managing the j3md gets out of hand. if you want to support multiple lighting techniques and stuff like that) My slang test bench currently allows adding techniques to all materials without having to declare them, declare the used interface specialization, automatic uniform / defines extraction. it does all that without any changes to the engine itself. it is basically only an assetloader. it should be possible to get the same featureset with a more magic glsl loader. but then you would need crosscompilation / reflection too. but just an experiment. not going to use it for now

zzuegg: > so i guess, with enough spare time slang is an option for the future we’d need a pure java implementation though, having a core functionality like this rely on bindings will limit engine portability. At some point graalvm’s webimage backend will be production ready and imo we should consider that as a potential target in the future