ODF is not open source, but it is an open standard

Print Friendly

I’m not sure if you’ve noticed this, but there’s been a curious conflation of the ideas of the open standard OpenDocument format (ODF) and the concept of open source. Now there has long been confusion between open standards and open source, and that’s why I wrote the four part series last year that teased apart the two. At some point, the statue of limitations for confusion will run out and those that can’t keep them straight will just look silly, or worse. In the meanwhile, we all have to keep educating people on the difference.

I’m hoping that it is only confusion when I see press articles about open document legislation referred to as “open source laws.” Obviously, they are nothing of the sort.

The particular confusion around ODF is more serious because I think it is intended to extend the operating system fight from Windows vs. Linux to Windows/OOXML vs. Linux/ODF. That is, if one had in place machinery to boost Windows and fault Linux, then tacking on the document standards is just the next step. Those of us in the ODF camp do not see it that way. ODF is useful on any operating system and in any application that wishes to support it. That is, it is multiplatform and can be implemented in proprietary or open source software.

As to OOXML, right now it is essentially Windows-only and Microsoft Office only and it remains to be seen whether a complete and full implementation can ever be done on another platform. Supporters will be under a tremendous amount of pressure to 1) prove that anyone other than Microsoft can fully implement it, either for technical or legal reasons, and 2) prove that there will be many applications on multiple platforms that support it. Otherwise, it will will just die away as I have previously predicted.

So let’s divorce open source from open standards. Choose your operating system, proprietary or open source, and then carefully choose your open document standard to allow fully interoperability and interchangeability of applications.

Addendum: This entry was originally titled “ODF is not open source,” but that seemed to confuse some people. I augmented the title to make it a bit clearer, I hope.


  1. If you want copies of IBM Lotus Notes, or Sun StarOffice, then you need to buy them from your friendly IBM salesman or Sun salesman, respectively. Or their business partners.
    If you want copies of OpenOffice.org, you can download them from http://openoffice.org , or maybe they’re for sale at a school fete to raise funds for the PTA, or maybe you can find then growing in the garden like daisies.

    All of them claim to implement ISO26300. What you do if you find they don’t implement ISO26300, or you’d like ISO26300 enhanced in some way, rather depends on which one you have, how you got it, and what your skills and budget are.

    If you want copies of Microsoft Office, see your Microsoft salesman or business partner, of course. Simple.

  2. I got onto the UK committee slogging its way through the ecma standard on OpenXML by mistake, but I seem to be stuck with it. Echoing what you say above, the thing that irritates me most is that I cannot see any reason for a second ISO/IEC standard in this area. MS claims OpenXML is better for the billions of legacy documents out there in old MS file formats. That would seem to boil down to:
    (1) all (or maybe almost all) legacy documents can be satisfactorily converted into OpenXML; and
    (2) a substantial (or significant) proportion of legacy documents cannot be satisfactorily converted into ODF.
    Both these propositions are essentially questions of fact. So the question is WHERE IS THE DATA? As a bare minimum, can we have a reasonable selection of legacy documents in support of (2) (so that they convert well into OpenXML, but not into ODF).
    I recently asked this question on Brian Jones’ blog and have not yet had any useful answers. Nor have the MS members of the committee come up with anything yet …

  3. John, imagine you’ve got a bunch of VHS videotapes with stuff on. They’re rather like these billions of MS ‘.doc’ documents.

    You’ve also got a few Betamax videotapes, which are rather like Lotus SmartSuite documents. ‘1-2-3’ spreadsheets and the like; famous, but history. Not many new ones of those are being created nowadays.

    You know that the future lies with ISO standard DVDs.

    Yes, converting the VHS videotapes to ISO-standard DVD probably will lose some of the details of the formatting.

    So the question facing the “UK committee” is whether it’s appropriate to have another ISO standard for DVDs which will let you faithfully transcribe the old videotapes; or whether you should keep a few VHS tape players (copies of Microsoft Word 2003) around in case you need to read them.

    Personally, I’d say to leave it at ISO26300, and not have another ISO standard. But that’s personal; I don’t speak for my employer (though I’m fairly sure I know what he would say); nor do I speak for any other business or individual. You’d have to ask them to speak for themselves.

  4. Chris, if you click my name to get to my blog on MS, you will see that I at least am reasonably sympathetic to your point of view … :)
    I think quite a few people involved in the UK decision are fairly angry with the JTC1 secretariat (which just happens to be ANSI) for the way they ignored the UK’s contradiction, which should have stopped the fast track process (also dealt with on my blog).
    On the other hand, others tend to think that since MS Office is the de facto standard with 97% of the market, we are all better off if control of OpenXML is taken away from MS and passes to an ISO/IEC committee …

  5. Well, it wouldn’t stay with ISO very long.

    In principle, IBM might want to come up with an ‘OOXML Understander’ to run on zSeries (mainframe) Websphere as part of an office automation solution for a well-funded client. So, taking that as an example of an application which doesn’t appear on a user’s desktop or accept their keystrokes, you can probably think of others.

    Firstly, no-one but Microsoft can sensibly implement the specification, nor can they test conformance. Previous articles on Bob’s blog estimate 150 person-years to implement, even if you make sensible guesses as to the imperfectly-specified portions. And then, how do you even begin to test conformance and interoperability ?
    So, why bother to spend 150 person-years of engineering effort, when there’s a perfectly good ISO spec there already, with several implementations on the market (and no-charge ones available for anyone to pick up from open project web sites) already ?

    Secondly, if Microsoft’s track record is anything to go by, they will steadily add more to the document spec in what appears to the rest of the world to be an ad-hoc fashion. That’s how we got from ‘Word 1.0’ back in 1985 or so, to Microsoft Office 2007 now; it’s how Microsoft persuade people to part with more money in the direction of Redmond with each new release. Yes, new variants are arguably better in various ways; but they achieve that at the expense of interoperability with old variants of the same product.

    There’s no reason to think they would change the habit of a lifetime. So, even if ISO do issue a standard, pretty soon there won’t be any implementations of the standard.

    Glad to hear you’re on examining the proposal. Don’t take any money privately from any corporation (not Microsoft, not IBM, not Sun, not any of them)… take salary from your employer, and if you take any money or rewards from anyone else, declare it publically to the UK committee. And give your own professional judgement when it comes to the vote.

    I think I know what’s best. But I’m not on the committee. I don’t get a say.

  6. Stephane Rodriguez

    “Secondly, if Microsoft’s track record is anything to go by, they will steadily add more to the document spec in what appears to the rest of the world to be an ad-hoc fashion. That’s how we got from ‘Word 1.0′ back in 1985 or so, to Microsoft Office 2007 now; it’s how Microsoft persuade people to part with more money in the direction of Redmond with each new release”

    It’s true, but it’s far worse than that. Back in the days, Office was a full client, there was no server-side. Beginning with Office 2007, Microsoft is beginning to integrate many server pieces together : sharepoint, excel services, sql server, …

    So what we’ll be talking about in the future is not just office file formats, it’s office file formats being part of a bigger picture which includes many more proprietary pieces (servers), and many more undocumented protocols (such as Office Word storing history of servers where it has published a Word document with a profile security restriction).

    To those thinking OOXML is the end of Microsoft’s disingenuous “open” initiatives, this is actually just the beginning. Just you wait they start adding .NET 3.0 pieces and its proprietary type system into it as well.

  7. It rather says that “Microsoft Office” is a product, or a solution; and things like XML specifications of storage formats are part of the product specification.

    It’s useful that they write down the product specification, to say how it does operate.

    But it’s pointless asking ISO to issue a standard for it; that would be a specification of how office productivity solutions should operate. And the question of how office productivity solutions should operate, is either none-of-ISO’s-business, or to be decided jointly by a committee of academics and many intending vendors in the ‘Information Technology’ business.

  8. Any one of you gentlemen knows what format is used by Google Documents?

  9. Google Documents states that you can upload the following types:


    • HTML files and plain text (.txt).
    • Microsoft Word (.doc), Rich Text (.rtf), OpenDocument Text (.odt) and StarOffice (.sxw).


    • Comma Separated Value (.csv).
    • Microsoft Excel (.xls) files and OpenDocument Spreadsheet (.ods).
  10. I would think IBM, Sun and the like would be only too willing to explain the difference between “open standard” and “open source” to respective governments in this way:

    You have a word processor – let’s say AbiWord, which is cross-platform, open source, and uses ODF – on the client PC. You have a network, which connects to the big iron. And on the big iron, you have a big iron server operating system, such as AIX, Solaris, or zOS, which is running a text server – for IBM, say, WebSphere. Now WebSphere can handle ODF, though it isn’t itself open source. Each use of ODF could be completely different – the word processor does text input and display; the text server collates selected elements from each document it receives, and hands the resulting statistics off to a proprietary database/server. Et cetera. Each is a valid use of the ODF file format.

    I think that would go a long way to explaining the difference between open source and open standards. Open source is a way of developing software that prevents lock-in for software developers and customers; open standards are a way of developing and maintaining a marketplace free for all comers.

Comments are closed