Share via


Open XML BRM Day 2 Thoughts

Well it would seem that facts, rumors, and opinions are flowing fast and hard from Geneva. There have been some substantive opinions being shared around that are interesting. I'll focus on 2 topics for today.

1) Patrick Durusau - the project editor for the OpenDocument Format has posted another opinion, and if you are serious about thinking deeply on this topic you should be reading what he has to say. He is a long-time standards professional and I find his writing to be extremely thought provoking. I have great respect for his approach.

  • Patrick makes the point that ODF and Open XML are fundamentally different in their approach. He, as many have, contemplates the idea that ODF and Open XML are markups which are evolving at the same time. He makes the point that it is encouraging to have both formats within the JTC 1 context as it will promote a balanced, technical conversation about the future of formats. I'd also point out that DIN - the National Standards Body for Germany has been driving a conversation on this. Organizations such as PC-Ware Information Technologies AG, Microsoft Deutschland GmbH, OPENLiMiT Holding AG, Dialogika GmbH, Ecma TC45, Novell and independent experts have been working to truly understand what the differences between the formats are. Here is the press release on that.
  • Patrick is very careful to state that he is not in favor or against the Open XML spec - my read is that he is in favor of the process for considering the spec in a holistic manner, rather than through the outright commercial and/or philosophical fanaticism that seems to be driving much of the opposition at this point (my opinion, not his).
  • Patrick points out that Open XMl and ODF are "...two of the major markup vocabularies.." I'll just underscore a theme that is raised all the time, that somehow ODF should be the only document format. There are many today (more than the two being discussed), there will be more in the future. I think it is very good to have them standardized and open for all to use. Great. Now let's move on to the real issues of translation.

2) Google has decided to join the OpenDocument Alliance in opposition to Open XML's bid to become a JTC 1 standard. They posted this opinion that ODF should be the only format and that national standards bodies should vote no to Open XML. They have brought out three primary arguments against Open XML - all three of which are existing opinions of other detractors. My two cents:

***AMENDED BASED ON COMMENTS: I mispoke in the opening of this section and then compounded it by simply missing the point of the org names that I had wrong. I originally wrote that Google was joining the OpenDocument Foundation. Google has been a member of the OpenDocument ALLIANCE for quite some time. The OpenDocument Foundation has long-since closed their doors and the folks over at Groklaw were right in hammering me about getting the names wrong. I have written at length about the Foundation as opposed to the Alliance and should have had this right. My apologies and thank you to Kevin Wright and Luc Bollen for their comments below.***

  • Google has every right to dislike the specification and advocate against it. It is worth pointing out that they have a commercial interest in their foray into the marketplace with Google Docs. (which is just fine, but should be kept in mind when viewing anyone's opinions...including my own.)
  • They start with what was the contradiction argument of more than a year ago - which ISO determined not to be the case. More than that, so many people have written on this point that no arguments are left other than to say simply...they are wrong. The formats represent both the past, present, and future of different applications and different functions. ODF represents a subset of capabilities offered in Open XML and does so due to the products that were the impetus for the creation of the formats. Never mind the fact that the derivation of the formats took entirely different engineering pathways and take into account different sets of legacy issues (for good or ill).
  • Their second point also misses the point. The idea that the specification represents the existing data of millions of organizations' data is not a flaw. I do hope that as they bring along their customers from feature to feature (technology to technology) over time, they are respectful of the need for legacy compatibility. The customers I speak to feel rather strongly about this. Additionally, now that there are literally hundreds of independent implementations on Linux, Windows, Palm, iPhone, etc....it seems a bit of a reach to still insist that the spec is limited to a single vendor.
  • Finally, they argue that changes in the spec demonstrate that it is "insufficient." Hmmm...the changes coming from ODF 1.0 to ODF 1.2 and 1.3 are huge - again, does that mean that ODF shouldn't have been standardized? I have written frequently about this. I have no problem with the fact that specifications change for the better. That is a good thing. Market relevance, the quality of spec for implementers, the terms governing use...all important factors. Let's focus on the real issues. Google doesn't want Open XML to pass because it creates commercial opportunity for them (as it would for IBM and/or Sun) - ok, I respect that. But the arguments are still thin as far as I'm concerned. (Probably not surprising to many that my opinion would differ.)

Comments

  • Anonymous
    February 26, 2008
    PingBack from http://www.mobiletechonline.info/?p=40385

  • Anonymous
    February 27, 2008
    The Microsoft spec is called "Open Office XML" (OOXML), not Open XML>  Calling it Open XML is misleading and disingenuous You should probably read Groklaw's corrections of your blog and correct your mistakes, such as Google joined the OpenDocument "Alliance" not Foundation. http://www.groklaw.net/article.php?story=20080227031234460

  • Anonymous
    February 27, 2008
    "the changes coming from ODF 1.0 to ODF 1.2 and 1.3 are huge" This is simply not true !  OASIS decided to standardise (as ODF 1.0) the core of the specification, which was mature.   As a result, not a single change of the text was needed as part of the ISO process. The work continued on the other parts (e.g. the math formulas), which are now stabilising and will be ADDED to the existing specs.  The existing specs will NOT CHANGE. This is very different from OOXML, which is so immature that the bulk of the review is taking part during the ISO process, and is resulting in VERY SIGNIFICANT CHANGES (not additions! - e.g. hundreds of pages being deprecated even before being standardised...) to the spec just at the end of the ISO process.  

  • Anonymous
    February 27, 2008
    Afraid to allow comments on you Blog?

  • Anonymous
    February 27, 2008
    The comment has been removed

  • Anonymous
    February 27, 2008
    "They start with what was the contradiction argument of more than a year ago - which ISO determined not to be the case. " Presumably, you meant  "An Unnecessary Standard". Whatever the case, I am not aware that ISO actually deliberated anything for OOXML. As far as I can tell, it is, by its own rules, obliged to give everyone (NB and ECMA) a fair hearing. While it is true that this issue had been raised by NBs early in the process, ISO had not explicitly said that this is not the case. It merely allows the next stage to proceed. In fact, on retrospect, when the NBs raised the issues on whether to accept OOXML for fast track, they are simply advisory and is used by ISO as an indication whether the fast track process is worthwhile. Looking at some of the comments from NB, the issue is not resolved as the question of ODF vs OOXML is still listed as comments. Surely if  the issue had been resolved, NBs will not raise this? "The idea that the specification represents the existing data of millions of organizations' data is not a flaw." Huh?? Millions of organization has data in MSOffice binary format . But OOXML? MSOffice sales figure will suggests that data in OOXML is a rare minority. Anyway their argument is that legacy stuff (those marked as deprecated) need not make it into OOXML as it can be handled internally by MSOffice. As for the "insufficient standard" argument, the first paragraph makes it clear that Google does not mean missing parts such as formulas etc, but the inability for implementors, such as Google themselves, to implement items that had been documented by OOXML. Compared to Google's earlier statements before the September vote, this statement is well-thought through. The September statement looks like recycling existing statements from other party. This statement shows that Google had studied things carefully.

  • Anonymous
    February 27, 2008
    Your arguments ring hollow. Let's not forget that Microsoft is a convicted monopolist both in the USA and the EU. That alone discredits anything you say here. Add to that, Microsoft's attempt to stack standards committees; Microsoft's undermining of government and business officials who decide that a non-Microsoft solution is best; Microsoft's FUD in buying "independent studies" to "prove" that Microsoft is best, and Microsoft's credibility is zero, zilch, null, void, nothing. So, in my shop, out of about 3200 seats, there are only about 50 running Microsoft anything. Those remaining seats will be converted to open-source equivalents by year end 2008. Have fun in your walled garden. Rex

  • Anonymous
    February 27, 2008
    > Well it would seem that facts, rumors, and opinions are flowing > fast and hard from Geneva. Regarding that, MSFT and ECMA could do away with a lot of the "rumors" by actually opening the process, instead of trying to conduct all proceedings and revew under a cloak of secrecy.  After MSFT's previous conduct in this affair was noted and called out for exactly what it was, it's actually quite reasonable to assume (or at least suspect) the worst: that such behavior is still ongoing. If MSFT is actually conducting itself better now, then letting public eyes peer in and verify that is the best thing you could do for the credibility of the process.  Cloaking the whole thing in secrecy implies that MSFT's tactics are still something to be ashamed of. > Finally, they argue that changes in the spec demonstrate that it > is "insufficient." Hmmm...the changes coming from ODF 1.0 > to ODF 1.2 and 1.3 are huge - again, does that mean that ODF > shouldn't have been standardized? I would say that the changes, coupled with the outrageous size of the spec and the inadequacy of the allotted review time, do indeed demonstrate that the spec is "insufficient."  If it was truly sufficient, then the brief, cursory, and hopelessly incomplete review it already received should have turned up far less problems (try zero).  Finding glaring flaws in only a small subset of a design suggests that many more (unknown, but equally critical) flaws exist in the rest of it.

  • Anonymous
    February 27, 2008
    The author of the blog post is either intentionally misleading (not unexpected of Microsoft) or else needs to check his facts.  He wrote: "Google has decided to join the OpenDocument Foundation" Whereas Google's blog post says "We join the ODF Alliance" Also, "join the ODF Alliance" means "join our voices", not "join the organization".  Google has been a member of the ODF Alliance since 2006.

  • Anonymous
    February 27, 2008
    @Kelledin "MSFT and ECMA could do away with a lot of the "rumors" by actually opening the process" The current "closed" nature of BRM and the fact that NBs' comments are made secret have to do with ISO procedures, not ECMA. As such, it will be wrong to blame ECMA and MS for this. I am agnostics about close door meetings at ECMA and/or ISO. Sometimes, closed meeting allows participants to express their opinions more freely.  However, in the interest of  transparency, closed meeting should be kept to a minimal w.r.t. the BRM process, ISO is clearly only interested in NBs' opinion and not joe public's. Hence, closed meeting is probably more appropriate. w.r.t . ECMA, while I am not saying that they should adopt the OASIS way, more transparency is needed.

  • Anonymous
    February 27, 2008
    @Kelledin If one read really carefully Google's statement on "Insufficient Standard", what it meant is there are insufficient information to implement certain aspect of OOXML standard. Google's beef is that OOXML tells it that it has to support functionA but do not give enough information to actually sit down and write functionA. This is different from what Matusow is sayiing. What he is saying is there are  functions left out in ODF, which is true. However, it is not what Google is saying.

  • Anonymous
    February 27, 2008
    Thanks all for the comments - Luc - the point is that specs improve and change in the process. I haven't done a detailed rip-down of the ODF spec, but I would be surprised to learn that nothing has changed in the core spec between 1.0 and 1.2. I open to being educated on this, but I find that assertion a bit suspect. No matter what, the point still remains, specs change and improve - that is built into the process and is not out of scope. Market adoption and the use of the spec in implementation remains the real test for any spec - and that we are clearly seeing with Open XML and ODF today. Dave - my guess is there are very specific commercial reasons for Google not wanting to support Open XML. Over time, if we continue to see the growth of adoption that we are with the format, the calculus used to determine their support of the format may well change. IBM is supporting Open XML in a few of their products - I'm sure that is not because they like the spec nor Microsoft. Customers though, that is a compelling reason. Wu MingShi - thanks for all the comments. Thoughtful as usual. There were arguments put forward that there was a "contradiction" between Open XML and ODF at the beginning of the process. My understanding is that  a decision is made at that point by JTC 1 based on those argument. If a contraction had existed then the process would have been halted. (Someone can correct me on this if you have more info on that part of the process). My point was that Google was rehashing an argument long-ago put to rest. There is absolutely no reason that prevents an app from using both formats at the same time. There is no contradiction. As for the millions of docs - you are right, I meant in binary. Having a design goal of providing for all the legacy docs to be able to move forward is NOT about making this spec single-vendor in focus. But it IS about respecting the legacy investments of organizations around the world. Thus, to dismiss a core design goal as they did in that letter is what I have a problem with. Given the independent implementations out there today of the Open XML spec...it would seem the single-vendor argument is not valid.

  • Jason
  • Anonymous
    February 27, 2008
    The comment has been removed

  • Anonymous
    February 27, 2008
    I've been following the OOXML thing for a while. I'll try and keep personal opinions out of this - I find Microsoft's business practices (lack of interoperability, monopolistic practices, etc.) to be... questionable. If this new interoperability promise rings true, it will be a great day for you guys, and maybe I'll touch a Microsoft product for the first time in 5 years. My personal biggest concern with OOXML is that it may be encumbered by patents. What's the purpose of a standard if it can't be fully implemented by others? And when I say fully, contrary to my reading of the latest interoperability announcement, I'm talking about it being implemented in GPL'd software, non-commercial or commercial, without any liability or licensing cost at all.

  • Anonymous
    February 27, 2008
    Luc Bollen wrote: "The existing (ODF) specs will NOT CHANGE." So, this means ODF will still not support SVG after all?

  • Anonymous
    February 27, 2008
    Jason to Dave: my guess is there are very specific commercial reasons for Google not wanting to support Open XML. Well, and then why doesn't Microsoft support ODF as a first class citizen? This would tell a lot about your will to interoperate? Stop throwing stones at the others if you live in a house of glass.

  • Anonymous
    February 27, 2008
    If you want to be taken seriously, you should try to sound less like one of the many Microsoft shills. Oh, and fact checking is not just a nice thing to do.  Other people will agree with me, when I say that I cannot take you seriously any more.  I mean, when exactly did Google join what ODF "Foundation" (not!), and did they really say that it was against OOXML? I mean, if you did fact-check what you wrote, you would have written something along the lines that OOXML is just not good enough to be a standard, i.e. something to serve all parties. Obviously it is not evident for everybody... Sigh.

  • Anonymous
    February 27, 2008
    here you are the Durusau OOXML "openess": http://europa.eu/rapid/pressReleasesAction.do?reference=IP/08/318&format=HTML&aged=0&language=EN&guiLanguage=en  marc

  • Anonymous
    February 27, 2008
    The comment has been removed

  • Anonymous
    February 27, 2008
    The comment has been removed

  • Anonymous
    February 27, 2008
    Jason, I find it always interesting to see the perceptions MS employees have of their company and the rest of the software world in general. While I can understand your point of view that there are standards "contaminated" by RAND clauses, I urge you to question what the landscape would be like if "necessary" standards were patent encumbered. Think TCP/IP or SMTP. I think I can safely say that any standard that actually aims for widespread adoption should be "required" not to need a RAND clause. I believe most companies that participate in standards discussions and which actually endorse RAND terms for a proposed standard are either eyeing a direct payback for the support or contemplating proposing a standard themselves that will make them rich when it's adopted later. Also since you do make the claim that there are many standards with RAND specs implemented under Linux, could you please care to enumerate them ? cheers. Faisal.

  • Anonymous
    February 27, 2008
    Luc Bollen wrote: "OASIS decided to standardise (as ODF 1.0) the core of the specification, which was mature.   As a result, not a single change of the text was needed as part of the ISO process. The work continued on the other parts (e.g. the math formulas), which are now stabilising and will be ADDED to the existing specs.  The existing specs will NOT CHANGE." That will disappoint at least the standardization bodies BSI, DIN, EOS, JISC, MSZT, SAC, SZZ and SII, as eight of the standardization bodies that voted 'approve' but with interesting and relevant comments on the core of ODF during the fast track process of ODF in 2006. Also, you may want to read chapters G3 and G4 in the ODF 1.1 specification in the section "Changes from Previous Specification Versions" and you may want to look for the Accessibility features in ODF 1.0 and the ODF Access Requirements report on the 1.0 version. But, then again, it's all beside the point, and not a real issue to debate about. All standards improve and change.

  • Anonymous
    February 27, 2008
    @jason "There were arguments put forward that there was a "contradiction" between Open XML and ODF at the beginning of the process. My understanding is that  a decision is made at that point by JTC 1 based on those argument." I assume we are both talking about the comments ISO solicited from NBs before it decided to go on to the 5 months period leading up to the September vote. I am not been involved in any standardization process. Perhaps we are using slightly different definition for "contradiction". I believe you are trying to say  that any contradiction, if detected, will automatically lead to the rejection of OOXML for fast track. I believe this is your argument and  I acknowledge that your argument may turn out to be the correct one. However, as I understand it, to reject OOXML based on contradiction at the stage can only be done if the proposal was frivolous or simple modification of an existing standard, and OOXML is not frivolous or is 99% a copy of an existing standard. Moreover, contradiction is not a "black-and-white" issue but rather comes in shades of grey. IMHO, to argue OOXML contradicts ODF falls in the "shades of grey" category and it is correct that ISO put it through the 5 months process for NBs to discuss this and other issues with OOXML to decide on whether to approve OOXML.

  • Anonymous
    February 28, 2008
    First An article in CIO Magazine online The CIO of Denmark comments that no matter what happens at the

  • Anonymous
    February 28, 2008
    hAL and Hans - thanks for the valuable input. Fasial - you have misunderstood my point about RAND. I do not believe that RAND "contaminates" - that is your word. RAND terms have grown out of more than 100 years of standardization legal precedent. RAND is a foundational element in bringing fairness to the standardization process. RAND is not the same from forum to forum, that is why each standards organization has its own IPR policy. RAND generally has many terms associated with it - only one of which is about royalties. One common term is about the limitation of scope for the patents associated with the implementation of a specification. The Free Software licenses are not compatible with that term. The choice of the FSF to create that term does not make the entire rest of the world incorrect for having used it. You wanted examples...how about Unicode for a start? There are many more. This is a complex issue that does not have a right or wrong, but does live within the context of existing law and precedent. Wu MingShi - thanks for the summarization. Yes, I think you have it about right. The macro point I have been making is that the argument that there should be only one standard, when predicated on the "contradiction" point, is not well founded. As someone who believes in market competition for standards (TCP/IP vs. OSI as the classic example), I do not agree with the whole premise of Open XML not being standardized just because ODF was brought to JTC 1 first. Thanks, Jason

  • Anonymous
    February 28, 2008
    First An article in CIO Magazine online The CIO of Denmark comments that no matter what happens at the

  • Anonymous
    February 28, 2008
    The comment has been removed

  • Anonymous
    February 28, 2008
    Jason, "AMENDED BASED ON COMMENTS: [...] Google has been a member of the OpenDocument Foundation for quite some time. [...] Thank you to Kevin Wright for his comment below. You missed the main point of Kevin's comment : Google has been part of the OpenDocument ALLIANCE for quite some time.  It has never been part of the OpenDocument FOUNDATION (which no longer exist, by the way).

  • Anonymous
    February 28, 2008
    The comment has been removed

  • Anonymous
    February 28, 2008
    The comment has been removed

  • Anonymous
    February 28, 2008
    Luc Bollen wrote: "The new text containing these huge changes will be available only at the end of March, a couple of days (if we are lucky) before the deadline for NBs to change their votes." I'm glad you can see into the future and tell when ECMA will have the new text available. Can you please tell us the outcome of the voting too?

  • Anonymous
    February 28, 2008
    @Frederik : The JTC1 rules is that the final DIS text has to be issued "not later than one month after the ballot resolution group meeting".  This is end of March. Now, indeed, the editor can rush and issue the text a couple of weeks earlier.  But I'm not sure that this would be a wise move, given the volume of change to be included, and the major reshuffle of the text recommended by ECMA in their dispositions.

  • Anonymous
    February 28, 2008
    @hAl: "Also if the OOXML ISO standardisation fails it would be interesting if Microsoft and a bunch of its partners will the join the OASIS TC for instance to amend the format to allow for backwards compatibility features. With the current level of control of the ODF spec by IBM and Sun that might lead to big problems in OASIS." Indeed, it would be very interesting, and personally I would applaud this move, provided this is done in good faith and not by stuffing the OASIS TC with shills. And it would reduce the level of control by IBM and Sun, which I would consider also as a very good thing : OASIS can become the open place were everybody talks to everybody, as recommended by Patrick Durusau.  

  • Anonymous
    February 28, 2008
    Speaking of standards, formats, and spreadsheets, I found out the other day that the Open Formula people have done something that's been needed for ages - a BNF grammar for speadsheets, in effect treating the spreadsheet as a specialized programming language (which it is) and standardizing it so that competition in that space will become easier.  I will trust engineers a lot more if they use spreadsheets that rely on a publically-defined BNF grammar.  A lot less automagicality, a lot more reproducibility. And Jason, since you mentioned the TCP/IP vs OSI networking standards, how many OSes these days use FTAM?  versus FTP?  The only time I've ever seen fragments of an OSI stack was when I looked inside 4.4BSD ... ODF isn't even the first file format standard either.  The first was ODA: http://en.wikipedia.org/wiki/Document_file_format http://en.wikipedia.org/wiki/Open_Document_Architecture It's just the first with momentum - and with Google behind it, it's going to take quite a lot more stopping than Microsoft has ever considered possible before.  And Microsoft's under the anti-trust microscope this time.

  • Anonymous
    March 26, 2008
    Why do you still call it Open XML? The proposed standard is not named that. That seems misleading.

  • Anonymous
    March 26, 2008
    The comment has been removed