Announcement

Collapse
No announcement yet.

Microsoft Open-Sources Its OOXML SDK

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

  • #11
    Originally posted by afeder View Post
    If that is the case, I guess Google are not in their right mind given that they have just announced native OOXML support in Google Docs. Regardless of how poor the format is, it is too widespread for real businesses like Google to ignore.
    I mean except big corporations of course since for example Microsoft a priori supports OOXML.
    As to Google I'm pretty sure it only supports a subset of OOXML since even ODF isn't supported by any toolkit 100%, last I read LibreOffice was adding minor missing ODF functionality, and as I read a while ago even Microsoft wasn't fully supporting OOXML since it's too much work.

    In Google's case one has to notice that there's many levels of support for something:
    1) partial support, even that partial part might be crappy
    2) full support, but crappy or not optimized
    3) full support, good and optimized

    Does Google state which type of support for OOXML they have?

    And yes, "native" basically doesn't mean anything. Remember how Microsoft praised IE9 for "native" hw acceleration?

    Comment


    • #12
      Originally posted by afeder View Post
      (OOXML itself is a hell, if ODF takes a little over a 1000 pages of documentation, OOXML is 5-10 thousand pages long. No one in his mind would bother to implement such a huge pile of prescriptions.)

      ..If that is the case..
      Yes it's the case, I actually implemented the core part of .ods (ODF spreadsheets) in Qt5 as a library for Qt users to be able to read/write .ods files, and the documentation for ODF is about 1200 pages, otoh OOXML (I looked it up out of curiosity) is out of 4 parts and the 1st one is the biggest one and is 5000+ pages.

      Comment


      • #13
        two things:
        1. Does it really matter how complicated the documentation is? It has obviously been shown to be a great file format... and this way, we can implement it without relying so much on documentation (and also improve compatibility with MS Office)
        2. Does this mean I can finally effing open macros created in MS Office in Libreoffice soon? I have a couple advanced budget spreadsheets that use MS Excel macros, but Libreoffice doesn't even notice it (forcing me to use Windows at least every payday, but usually more)

        Comment


        • #14
          Originally posted by Daktyl198 View Post
          two things:
          1. Does it really matter how complicated the documentation is? It has obviously been shown to be a great file format... and this way, we can implement it without relying so much on documentation (and also improve compatibility with MS Office)
          2. Does this mean I can finally effing open macros created in MS Office in Libreoffice soon? I have a couple advanced budget spreadsheets that use MS Excel macros, but Libreoffice doesn't even notice it (forcing me to use Windows at least every payday, but usually more)
          1. Yes. It really matters, just like it matters if you have to do 30 or 150 push ups. And no, it's just a modern file format, nothing special or great about OOXML, just XML with lots of rules, just like ODF.

          2. 90% of the time you use 10% of the features, iow if you use advanced OOXML stuff on non-ms-office you're likely to get problems, no consolations here. And don't rely on LibreOffice supporting OOXML (as well as ms office) any time soon. Just don't waste your time on hoping. We've been hoping here for Wayland for 6 years and we're still not using the fucker. There's sooo many missed deadlines, so just don't expect stuff to happen soon, just use ms office.

          Comment


          • #15
            Originally posted by mark45 View Post
            1. Yes. It really matters, just like it matters if you have to do 30 or 150 push ups. And no, it's just a modern file format, nothing special or great about OOXML, just XML with lots of rules, just like ODF.

            2. 90% of the time you use 10% of the features, iow if you use advanced OOXML stuff on non-ms-office you're likely to get problems, no consolations here. And don't rely on LibreOffice supporting OOXML (as well as ms office) any time soon. Just don't waste your time on hoping. We've been hoping here for Wayland for 6 years and we're still not using the fucker. There's sooo many missed deadlines, so just don't expect stuff to happen soon, just use ms office.
            As I remember, a lot of that bulk in the OOXML spec is duplicated effort (eg. defining dates differently in different parts of the spec) and legacy bodges accumulated in the binary formats that Microsoft decided to keep.

            Comment


            • #16
              Originally posted by mark45 View Post
              1. Yes. It really matters, just like it matters if you have to do 30 or 150 push ups. And no, it's just a modern file format, nothing special or great about OOXML, just XML with lots of rules, just like ODF.

              2. 90% of the time you use 10% of the features, iow if you use advanced OOXML stuff on non-ms-office you're likely to get problems, no consolations here. And don't rely on LibreOffice supporting OOXML (as well as ms office) any time soon. Just don't waste your time on hoping. We've been hoping here for Wayland for 6 years and we're still not using the fucker. There's sooo many missed deadlines, so just don't expect stuff to happen soon, just use ms office.
              1. I should have clarified. I meant "Does it really matter how complicated the documentation is (now that we have a complete OSS implementation that we can use/base off of)". Obviously, if documentation was all we had to go off of, it would suck. Also, a file type can be good or bad based on the rules placed on the XML. It can lead to different file sizes, opening/rendering speed, etc. So far, OOXML has beat ODF in both size and speed, though it may just be the implementation/optimizations that go with MS Office.

              2. (((((

              Honestly, I prefer MS Office, but I can't use that under Linux (yet). So I'm forced to use (or try, at least) LibreOffice/Google Docs. As mentioned above, it doesn't always work out... but LO has definitely proven to be a great project.

              P.S. The "not using Wayland" thing is because until Mir showed up, nobody cared. As soon as Mir showed up, look how fast people cared (and how fast crap got done). Unfortunately, nobody started porting their DE to Wayland (or even de-X11-ing it) until Wayland had hit version 1.0 (despite there being several "stable" parts of the API ready before then). Don't blame the main Wayland devs, it really isn't their fault. It's mostly the community's fault in this specific case...

              Comment


              • #17
                Latest Linux News? Oh, LOL. Just LOL.

                Is it really "Latest Linux News"? I fail to see how some MS crap related to Linux. Not to mention OOXML is known for huge (about 6000 pages) and poorly written specs, totally screwed up. Up to degree where MS Office can't even parse simple examples created using MS specs. Should I mention such format should die by horrible death? OpenOffice document formats are far more sane in this regard. Though MS shown poor standards compliance in their office suite in this regard as well.

                But whatever, I fail to understand how this crap fits "Linux news". If someone wants good laugh on MS, Linux and so on, MS has released new Android based )) phone under Nokia brand. So, beyound your wildest imaginations, MS Linux has finally arrived in form of Android! Lol, wtf .

                Comment


                • #18
                  Originally posted by Daktyl198 View Post
                  Honestly, I prefer MS Office, but I can't use that under Linux (yet).
                  I wouldn't recomment MS products for sure. These are fairly unreliable guys who would backstab at convenient moment. There was IE 5.0 for *nix. But once *nix-based faced strong competition from Windows, MS just stopped releasing new IE versions for *nix to backstab *nix systems during competition and put Windows into advantage. Are you sure you want to fall into well known pit for another time? Some people seems to have short memories for sure. Yes, I mean MSOffice would go away from you solely at MS discretion and you can't do anything about it.
                  Last edited by 0xBADCODE; 25 June 2014, 11:50 PM.

                  Comment


                  • #19
                    Originally posted by Daktyl198 View Post
                    P.S. The "not using Wayland" thing is because until Mir showed up, nobody cared. As soon as Mir showed up, look how fast people cared (and how fast crap got done). Unfortunately, nobody started porting their DE to Wayland (or even de-X11-ing it) until Wayland had hit version 1.0 (despite there being several "stable" parts of the API ready before then). Don't blame the main Wayland devs, it really isn't their fault. It's mostly the community's fault in this specific case...
                    just fixing your point here i agree with the rest of it(i don't love Ms Office tho), this a super common misconception that probably has been explained a 1000 times around but well here we go for the 1001.

                    Mir is completely unrelated to wayland development speed now and in the past and since its announcement it hasn't fastened wayland adoption times at all, it just got more publicity, canonical just waited until wayland was stable enough to announce Mir and provide the sensation that canonical was pushing wayland forward nothing else(they specially waited for wayland to solve the really hard problems so they can just code easy parts without the heavy lift[kinda bited them back tho]).

                    the first problem you have to understand is that wayland on itself is the very basic trivial part of the massive problem that implies replace a graphic server not the hardest part, first of all let mention the groundwork needed to make something like Mir or wayland even possible in theory:

                    1.) kernel level massive rework in the KMS and GPU specific DRM backends for stuff like proper and future proof memory allocation+FD, render nodes, VM, GPU page handling, cheap page flipping, smarter vblank sync, DPMS, generic Input, multiple render targets, multiple render outputs, etc. (check logs in kernel cgit all the way to 3.0 kernel for GPU drivers), without this either wayland or Mir will be crazy talk and just this took years and is not 100% done(this here is really really hard, probably less than 100 devs in the whole planet can work at this level efficiently).

                    2.) EGL / OpenGL ES has to get done because GLX was a massive everest size road block if you ever wanted to be taken seriously talking people into wayland or Mir, of course this included kernel rework, mesa huge rework, shader compilers backends rework, buffer handling rework, memory allocation rework, massive cleanups, Xorg huge rework,etc. probably close to half of all the massive kernel rework in 1

                    3.) Generic X less input system was required since XInput was set number 1 most horrid piece of software ever written by the human race(prolly only second to XFont server) and it was absolutely unusable for wayland or mir in any shape, this included kernel rework(evdev), Xorg rework(was a massive code chop in several releases), etc. this one took a lot of time and thinking too

                    4.) To have any hope of decent compatibility for legacy X apps millions of LoC, endless hours of cleanups, massive reworks in many subsystems, several aproaches had to be done inside Xorg for XWayland to even have a fighting chance and loooooot of work has been needed since then to get it to its current state and still is not at 100%

                    5.) Massive mesa rework all the way from core mesa to backend specific shader compilers to integrate nicely all those previous changes, including nasty rewrites of shaders compiler logic (or waiting for the tool to be able to handle it like LLVM), add of new extensions like EGL_EXT_buffer_age or EGL_EXT_swap_buffers_with_damage, etc.

                    6.) Graphic servers touch everything even stuff that you don't expect it like looots of CLI programs, specially in the case of a monster semi OS like X was and this is probably the biggest time consuming part, you have to remove millions and millions of LOC in several thousand apps to isolate/rework/optimize/improve all the X code to even be able to launch an app without X and dont get a nice SIGSEGV and from there you have to rethink and reimplement all this millions of LOC to use the new system/techniques/features because you now have the possiblity of doing things that were absolutely impossible in X, etc

                    7.) Lastly, you now focus on the new graphic system design and polish the rough corners in 1-6

                    now after all this work is when either Wayland or Mir can be barely seriously bringed as a discussion topic("coincidentaly" this is the time frame where Mir was announced, ofc the contribution from canonical to either previous point was an astonishing 0 LoC, you just have to love their PR department), so today i think we got around 80% of 1-5 and around 30% of 6 is done, so we can now have something to show and experiment with in 7, so as you see even if mark danced naked in an Apple WWDC and carved Wayland/Mir in Steve jobs chest with an scalpel while grabing steve balmer long lost hair step 1 to 6 won't move any faster but sure as hell will get an awful lot of publicity to either party(again you got to love canonical PR department), so if you never followed the development process and suddenly you get super bombarded with publicity you will probably will think your statement is true but in LoC measurable reality was just nil, Ofc this was helped for many irresponsible devs that ran to scream wayland full support for next version and coudn't deliver because the infrastructure wasn't there yet.

                    bit of a long post but hope it helps to clarify some misconceptions

                    Note: non native english speaker here so forgive any mistake
                    Note: im not an Graphic stack DEV and i didnt took part in the dev process, so is possible i misplaced some info here and there, always crosscheck the respective project logs to be sure, i just tried my best in following the dev process because i found it fascinating so dont take this religiously
                    Note: for more in deep info you have many rockstars here like keith, marek, bridgman, tom, jayden, arseri, etc. ask them and praise their really hard work

                    Comment


                    • #20
                      Originally posted by jrch2k8 View Post
                      ...
                      That's the first time I've heard much of that, actually.

                      I knew it required a lot of infrastructure change, I just figured many people were being lazy about it until a "competitor" showed up (Mir, though I've known for a LONG time that the two's development were not related in any way). Certainly, the major DEs weren't exactly doing their part to port anything over to Wayland, even after the publicity ramped up; not for a little while, at least. At least, that's the way I saw it.

                      Comment

                      Working...
                      X