Australia and OOXML

Somewhere too early


There have been some poor decisions of late in Australia. Not playing Hauritz and persisting too long with the out-of-form Clarke and Ponting probably cost Australia the Ashes and has led to terrible self-flagelation. While it’s generally not done to take pleasure in the discomfort of others, I do think an exception can be made in the case of the Australian cricket team.

From various recent blogs and tweets I’ve noticed a fuss surrounding the decision by the Australian Government Information Management Office (AGIMO) to recommend the use of OOXML as a document format, and from the tenor of the comments it would seem this is being treated as similar calamity for Australia. However, there appears to be some misunderstanding and misinformation flying around which is worth a comment …

Leaving aside the merits of the decision itself, one particular theme in the commentary is that AGIMO have somehow picked a “non-ISO” version of OOXML. I can’t find any evidence of this. By specifying Ecma 376 without an edition number the convention is that the latest version of that standard is intended; and though I do think there is a danger of over-reading this particular citation, the current version of Ecma 376 is the second edition, which is the version of OOXML that was approved by ISO and IEC members in April 2008. The Ecma and ISO/IEC versions are in lock-step, with the Ecma text only ever mirroring the ISO/IEC text. And although (as now) there are inevitably some bureaucratic and administrative delays in the Ecma version rolling in all changes made in JTC 1 prior to publication, to cite one is, effectively, equivalent to citing the other.

[UPDATE: John Sheridan from AGIMO comments below that Ecma 376 1st Edition was intended, and I respond]

Comments (8) -

  • Jesper Lund Stocholm

    1/20/2011 11:08:11 AM | Reply

    Hi Alex,

    Is there a reference to the details of the decision in Australia?

  • John Sheridan

    1/21/2011 4:46:48 AM | Reply

    Thanks for this measured response Alex. I'm sorry to disappoint you but we are talking about ECMA-376 Ed 1. See why at



    • Alex

      1/21/2011 7:24:12 AM | Reply

      John hi

      Interesting stuff. If the first edition is intended, then ideally it should be specified so suppliers don't assume (as I did) that you mean the latest edition.

      However, looking at the document you linked to I think the decision to use Ecma 376 1st edition is based on a false premise. The statement there,

      "[t]he Office 2007 format is based on the ECMA-376 1st edition and the Office 2010 default format is based on the ISO/IEC 29500 “transitional” standard"

      is incorrect. Ecma 376 1st edition contains a number (dozens) of substantial deviations from the file formats used by Microsoft Office (as welll as hundreds of less serious errors) and these have only been noticed and corrected in the International standardization work over the last two years or so.

      If the intention is to specify a standard that corresponds to what MS Office 2007 actually does, probably the best way to specify this is by citing "ISO/IEC 29500-4" since this, the "transitional" version is formally scoped to correspond to what Microsoft Office actually did (the "existing corpus of legacy documents"), and will always be closing in on that goal. By giving an undated reference you always bind to the latest version and - right now - this includes sets of corrigenda and amendments which bring the Standard into much closer alignment with MS Office files than any of the preceding Ecma or International standards.

      • Istvan Sebestyen

        1/24/2011 1:59:56 PM | Reply


        your observation that usally the latest Ecma Edition counts is generally true, but not always. Sometimes we get from users / members the request to keep more than one Edition alive. This is the case e.g. in ECMA-379, but also in ECMA-376 (OOXML). We are keeping there for the time being both Edition 1 (the original Ecma Version) and Edition 2 (the ISO/IEC JTC1 Version approved) version alive. We will keep both alive as long as our member see rational behind it. Actually, in terms of downloading from our website, in 2010 at least it appears that Edition 1 is being dowloaded slightly more often than Edition 2 (this is difficult to say how to count, because in both cases we have several parts and in several presentation formats...). Anyway, any further changes in the course of the maintenance will be done - as always - in synch with JTC1 (which means about the same time basically, or with minimum delay).

        Kind regards,
        Istvan Sebestyen
        Ecma International

        • Alex

          1/25/2011 1:50:24 PM | Reply

          Istvan hi

          Certainly it can be useful to keep old editions of standards available. But in Ecma's case, if somebody cites Ecma xxx, and there are three editions available, which edition is meant?

          Is it like an undated reference in ISO, which binds to the latest available version?

          • Istvan Sebestyen

            3/8/2011 9:37:19 PM | Reply


            If there is only one, the latest edition of the standard is in force, then just ECMA-xxx is enough. In case there are more than one Edition alive then there is no default, but has to be specified which edition is meant.


  • Istvan Sebestyen

    1/25/2011 3:21:44 PM | Reply


    if you go to the Ecma public website to the living and withdrawn standards:

    there you find the list of living and withdrawn standards (incl. the download of the standards).

    For ECMA-376 you find that at the moment two living standard texts exist (That is very seldom at Ecma, but we are flexible enough to offer this, if members and users want...):

    So, there you see that currently we simply offering the two editions (as long as the Industry tells us to do so), and users can simply choose what to pick. On the TC45 activity page we clearly explain how those Editions are developing, but it is up to the user to select what Edition he wants to have. What is important in this case that you can not simply refer to ECMA-376, but you MUST add what Edition you are talking about. So there is no default. Of course this is a very special case, and the normal case is what you describe, that you have only one Ecma Standard Edition "alive" and that is the latest one. Then the old Edition moves automatically to the withdrawn standards (where you can still download). So this is also the most common case in Ecma.

    The reason why we are doing this is clearly that there are always a few such exceptional cases when you have to keep two edition in parallel alive (but only for a given periode of time). Here I think we are more flexible then e.g,. ISO (they simply do not allow something like that). About JTC1 I am not sure, if you can have two Editions at the same time that are both "alive".

    Hope I could explain this clear enough.

    Take care,