Announcement

Collapse
No announcement yet.

We Might Never See A New OpenGL Version, At Least Not For A Long Time

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • We Might Never See A New OpenGL Version, At Least Not For A Long Time

    Phoronix: We Might Never See A New OpenGL Version, At Least Not For A Long Time

    During past Khronos press briefings about OpenGL/Vulkan and in other communications, while Vulkan is the organization's big graphics API focus, it was implied during these conversations that OpenGL would continue to march to its own beat and evolve as needed. While OpenGL continues to be significantly used by cross-platform graphics application/game developers, it turns out there might not be a new official version for a long time - if ever...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Why would there one? I always questioned this "Mesa will reach complete OpenGL support with 4.5 but we know it won't be for long". Why there would be OpenGL 4.6? And if there wouldn't be one, would it be a bad thing?

    Comment


    • #3
      opengl evolves by intoducing extensions, not by declaring new version numbers

      Comment


      • #4
        Originally posted by pal666 View Post
        opengl evolves by intoducing extensions, not by declaring new version numbers
        Eventually you have so many extraneous tacked on extensions that not rolling them into one version number to easily check for guaranteed support on all of them is going to be a PITA.

        Comment


        • #5
          Originally posted by zanny View Post
          Eventually you have so many extraneous tacked on extensions that not rolling them into one version number to easily check for guaranteed support on all of them is going to be a PITA.
          checking support for unused extensions is a pita for driver developers

          Comment


          • #6
            They first have to see if the new "close to the hardware" APIs actually become widely used and successful outside of gaming consoles. At this point there are only a handful of Vulkan/DX12/Metal games/applications and almost nothing uses them exclusively. In time they may even fail because they require a lot more work and knowledge from developers, which also cost development time and money. In that case, OGL can rise again and evolve towards whatever is needed.

            Comment


            • #7
              Extensions are a great way to overcome the design limitations of the "fully programmable" OpenGL pipeline and maximize performance, however they have one ginormous flaw - they too often tend to be vendor specific and non portable, meaning that if you are not one of those developers, eager to get paid extra to optimize your product in the favor of a specific vendor who has money to spare to hamper its competitors (cough nvidia cough) you are phycked into doing a looot of extra work to get things to perform at their peak across multiple vendor hardware. Such developers are as rare as a hen's teeth...

              Quite frankly, there is no longer need for further OpenGL version, OpenGL is an inflexible, outdated dinosaur and its only course is extinction, that will happen gradually but soon. Writing new OpenGL code today already makes little sense, writing OpenGL code in 5 years will be stupid to say the least.

              Vulkan, being a low level API is all that is needed really. Sure, it is more complex to use, it is more verbose to use, but hey, nobody is preventing people from making high level abstractions on top of it, implementing whatever ready to use functionality that will, unlike GL extensions, work across the board.

              Comment


              • #8
                No need for versions, who supports more extensions is da winner

                Comment


                • #9
                  Originally posted by ddriver View Post
                  Vulkan, being a low level API is all that is needed really. Sure, it is more complex to use, it is more verbose to use, but hey, nobody is preventing people from making high level abstractions on top of it, implementing whatever ready to use functionality that will, unlike GL extensions, work across the board.
                  Slapping high level abstractions on top of Vulkan will get you back to OpenGL. Contrary to your belief, ease of use is what makes or breaks a framework. Otherwise the whole world would be programming in C with the occasional ASM routine here and there.

                  Comment


                  • #10
                    I would like to see a new OpenGL version that mandates all of the AZDO extensions already available.

                    Comment

                    Working...
                    X