Your survey is broken. I tried entering other type of developer, none of the others are remotely appropriate, and it refused to submit until I took one.
Bad survey techniques are the road to product failure. Bad data is worse than no data.
IMO, for Vulkan to succeed you need at least one higher level layer on top of what you have. There arenāt enough developers with the background, time, motivation and perseverance to really apply your work.
That would be a shame, because it rocks!
Iām getting frame rates over 250 compared to 70.
But the development cost of getting there is immense.
Iāve gotten the starts of a simple pipeline builder working, but now Iām blocked because I need a layout package like html/css/svg. Iāve been fighting for hours just getting FreeType hooked in.
Never underestimate the pain and complexity of 2D. Itās not glamorous, but itās probably 99% of the actual use cases.
Also, never forget that some of us write code and some of us assemble code. Iām an author. I can bang out a spatial search tree from scratch in a day or two, with tests. But it takes me over two days to get each one of the different cmake builds just to compile and 3/4 of the time I abandon a package because it wonāt build per instructions.
My brain is full of other stuff, no time or room to master make, cmake, gnumake, ant, VS, gradle and the plethora of other build systems. None of which, ever work directly per instructions, especially on Windows - where Iām stuck.
Iām sorry you had difficulty entering information. I just tried entering information into the āotherā field on the type of developer information and it worked fine. Iām using Google Chrome. Can you tell me what browser youāre using? I can then relay that to the author of the survey.
Iām boycotting Alphabet products and services to the maximum extent possible.
I strongly suggest you add an open ended āWhat do you feel we need?ā question in addition to the assumed answers.
Also Firefox here. A radio button is missing in all the āI willā columns.
I feel the questions are bit loaded. On one hand I am like āsure, I would take all of the aboveā. So I am making a prediction that this would be the result of the survey; who would even say ānoā to more free stuff. On the other hand I would want Khronos to focus on their core competency and make things on engineering rationale (not on popular vote), and not engage in some development adventurism.
One of your survey questions was about āshould there be some sort of tutorial/getting started app.ā I canāt remember the exact question.
Iāve been working on VulkanQuickStart for over a year (on and off). Itās usable now, Iām using it on two other projects. Itās a decent āframeworkā similar to MFC/WPF/QT etc. targeted at Vulkan. It provides a user modifiable pipeline, and a lot more, written c++.
Many of us really hated to see the OGL 2 default pipeline go away. Obviously it was grossly insufficient for high end work, but it sure gave you the ability to get something up and running in a hurry. For dedicated graphics developers, they usually have a bunch of frameworks lying around from other projects so they can just repurpose one of those.
Not so for me and a lot of others.
https://github.com/btipton-ds/VulkanQuickStart might be helpful. Whether users use it as is, or grossly modify it for their own purposes - it saves writing a lot of code before you can draw something useful.