Announcement

Collapse
No announcement yet.

WebP Causing Controversy In The FreeDesktop.org Camp

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

  • WebP Causing Controversy In The FreeDesktop.org Camp

    Phoronix: WebP Causing Controversy In The FreeDesktop.org Camp

    There's some upset individuals over the "image/webp" MIME type not being added to the FreeDesktop.org XDG shared-mime-info package for Google's WebP image format...

    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
    What, IANA might be so delusional that despite the de facto usage of "image/webp" they might associate a different mime?
    If so, maybe it's time to stop following that authority. Don't follow morons.

    Comment


    • #3
      Has Something Changed?

      I was under the impression that */x-* and */vnd-* were for non-IANA MIME types and that this is how they've been used for years. Am I misreading the established precedent, or is there an argument for exception?

      Comment


      • #4
        Originally posted by mark45 View Post
        What, IANA might be so delusional that despite the de facto usage of "image/webp" they might associate a different mime?
        If so, maybe it's time to stop following that authority. Don't follow morons.
        B.S. on this claim. Defacto usage is JPEG/JPG, PNG, GIF, for the web. Then there is TIFF/RAW [types of RAW assumed], for publishing, compositing, etc. Growing inclusion of WebP into software supporting the format does not make it a de facto usage type.

        Comment


        • #5
          This issue aside, I'd really like to see WebP supported in more browsers. There have been times I've needed alpha transparency and PNG size has been large whatever I've tried to reduce it while WebP was far more reasonable.

          Handy compatibility chart is here: http://caniuse.com/webp

          Comment


          • #6
            Originally posted by Marc Driftmeyer View Post
            B.S. on this claim. Defacto usage is JPEG/JPG, PNG, GIF, for the web. Then there is TIFF/RAW [types of RAW assumed], for publishing, compositing, etc. Growing inclusion of WebP into software supporting the format does not make it a de facto usage type.
            He didn't claim that webP is the de-facto image format for any purpose. He just said that "image/webp" is the de-facto mime-type used for webP images.

            Comment


            • #7
              Originally posted by Marc Driftmeyer View Post
              B.S. on this claim. Defacto usage is JPEG/JPG, PNG, GIF, for the web. Then there is TIFF/RAW [types of RAW assumed], for publishing, compositing, etc. Growing inclusion of WebP into software supporting the format does not make it a de facto usage type.
              Clearly you have no clue what the hell a mime type is.

              Comment


              • #8
                Indeed they should follow standards, but then again, standards are often push in by the existence of non-standard features, as we have seen in many programming languages. Btw what takes the IANA so long to give WebP an official MIME type

                Comment


                • #9
                  I do not see any information that Google has requested the registration of the image/webp format (or their webm format which has a similar non-standard usage) despite it being created by Google in 2010.

                  According to the bug report, the shared-mime-data maintainers want:
                  1. image/x-webp to be specified as the *main* mime type;
                  2. image/webp to be specified as an alias of image/x-webp.

                  The rationale for this is that image/webp has not been registered and mime types without x-* or vnd-* are not used as main mime types unless they are registered by IANA.

                  This means that in terms of shared-mime-data, both image/x-webp and image/webp will work as expected.

                  The objection from the people proposing the addition is that image/webp should be the main mime type, even though it has not been registered by IANA.

                  The rationale from these people is that image/webp has been used by other software and that using image/x-webp (which only occurs in the bug) would break that software. Software using shared-mime-data needing to resolve aliases, so would recognise webp files. However, downstream clients (e.g. clients of webservers using shared-mime-data) will see image/x-webp.

                  The shared-mime-data maintainers do not want to add an exception to their policy for adding new mime types just because clients are using an unregistered mime type that does not follow the mime standard, and that following the standard will break these applications that are not following the standard. Why should the webp mime type be the exception? What about other unregistered formats? If webp is added unregistered, it will set a precedence. The next unregistered type will say "you added image/webp, so why not video/h265?" (or whatever the mime type and format are).

                  Note that this is not just freedesktop requiring/requesting a mimetype registration:
                  * firefox (https://groups.google.com/a/webmproj...ss/Ds94Fxpmdck)
                  * python mimetypes.py (http://bugs.python.org/issue11362) -- even though the FDO bug mentions that Python uses 'image/webp', that mimetype is not in the mimetypes.py file (see e.g. /usr/lib/python3.3/mimetypes.py)

                  Also note that on http://lists.w3.org/Archives/Public/...1Nov/0079.html image/webp and audio/webm are the only unregistered mimetypes not using the standard x-* form.

                  Comment


                  • #10
                    Originally posted by Serge View Post
                    I was under the impression that */x-* and */vnd-* were for non-IANA MIME types and that this is how they've been used for years. Am I misreading the established precedent, or is there an argument for exception?
                    Exactly.

                    Since it's not IANA-registered, FDO can only add image/x-webp. Of course nobody needs that, since google decided to just use image/webp in chrome, and other projects already followed suit.

                    The real culprit isn't FDO, but google, who decided to use it without registering. The solution is not to patch yet another project with an inofficial mime type, but for google to finally get the type registered. Unfortunately the bug report doesn't mention whether or not that's in progress.

                    Comment

                    Working...
                    X