The podcast is available at:
https://www.computerworld.dk/art/39318
The interview was conducted in San Francisco in May, 2007 during the JavaOne-developer Conference.
Intro:
My name is Simon Phipps I’m Sun’s Chief Open Source Officer. I’ve been closely involved in, I actually live in Europe, so I’ve been closely involved in the debate over both the software patents where I helped form a grouping in the European Parliament that helped defeat the patent directive and I’ve also been involved extensively in the discussion over ODF and Microsoft’s Office 12 file format that they’ve given to ECMA.
00:30 - Question (Kurt Westh Nielsen).
Ok, and presently there is a lot of debate in Denmark going on, as whether the Public sector in Denmark should actually be forced to accept both the ODF format and Microsoft’s Open XML format - Why do you think that MS actually decided to open the format?
00:51 - Answer
Well there’s two different dimensions to that , first of all Microsoft has positively and genuinely been moving over to an XML based strategy with Office, which is laudable , and you know they took on Jean Paoli who was one of the authors of the XML spec and they have been doing fine work in heading in that direction.
As is characteristic of Microsoft their world view extends only to their own products, and consequently when they talk of interoperability for example, which is one of their current favourite phrases.
What they actually mean is they have built a tunnel that lets you into their world, and so.
When you look at what’s going on here, I believe they were headed in the direction of XML anyway for office, but I think that when they saw that actually a standards body at oasis was creating a genuinely open format with the design point of not building a one way tunnel into their world, I think that they were faced with the challenge of accelerating their plans and rushing through the xml file format for office 12.
They actually did have the opportunity in 2002 – they’re a member of Oasis – they could have joined the working group that did open document format, but unfortunately they, to use a curt phrase, they ‘flipped it off’ and decided they weren’t going to participate and I think they deeply regret it.
And I think it has caused all of lot a great deal of expense, and unnecessary debate, because they could easily have joined the discussion at that point, but as is their way they decided to produce a specification with an apparently identical design point but designed as a one way tunnel into Microsoft, whereas ODF is designed to be a base line file format to be shared by many applications.
02:36 - Question
So l mean looking at the specification now at hand for Open XML how would you sort of characterise that format, is it a real open format?
2:47 - Answer
Well there’s open and there’s open.
You know, if by open you mean is it a fully documented format – then yes it is, for goodness sake!
It comes in two and a half boxes of photocopier paper 6000 pages long.
It has got a description of everything it does so it’s open in that sense – whether it’s open in the sense that it is truly implementable by any developer on any platform I think that’s a completely different question.
There are lots of things in their specification that make the assumption that you are running on windows.
And to replicate the functionality of those data structures you have to do a great deal of background reimplementation of basic ideas in Windows, if you are not on the windows platform.
So if you define open as meaning ‘is there documentation available’? well, there largely is, there are a few places in the spec where the documentation is inadequate for an implementation, places that are there basically for backwards compatibility with old versions of Microsoft Office.
But the much more concerning features and the ones where the philosophy of the data structures is a philosophy that assumes that Windows is providing you with particular data structures pre-cooked.
And if you are not on Windows producing those data structures in a way that is compatible with windows , particularly with windows vista and all its DRM nasties is going to be incredibly difficult for an external developer.
So I don’t believe it is open in the sense, that anyone on Linux or on Solaris would be able to implement it.
And the case in point there is to look at Microsoft’s own difficulties producing that file format on the MAC, where their MAC implementation is way way way behind the Microsoft implementation because the challenges of all the ground up development work needed to support it are so great.
So documented? Yes.
Open? Well not open if what you mean is ‘Can it be implemented with equal effort everywhere?’ No it can’t be’.
It’s specifically designed to be easy to implement on windows and I’m not privy to their plans but it wouldn’t surprise me at all to find it was specifically designed to be hard to implement everywhere else.
4:58 – Question
So referring to the actual debate I Denmark, I mean in your view does it actually make sense to have the Public Sector accepting two competing open formats then, the ODF and the Open XML?
5:14 – Answer
Well I mean it’s obviously a ridiculous thing to do, but it’s also something that is obviously a pragmatic necessity for the Public Sector in Denmark, because the Danish market is so locked into Microsoft Office already.
To miss out on the format that they are using already is a very difficult thing to do, but on the other hand to be able to have the flexibility in Denmark to have a choice of vendors throughout the infrastructure and desktop that you are using, you really do have to send a clear message to Microsoft, that they are not the only vendor you are going to work with.
Now you know Microsoft has made some tactical investments in Denmark, bought companies, become a powerful lobbying force, played standards politics and so.
You know you can’t ignore them over there, but you know you’re a freedom loving country just like Britain is and neither of us likes being bullied and Microsoft is being a bully and consequently ODF is actually the format that embodies Danish basic principles.
So so I think that pragmatically you have to have both.
I think if you are going to have Office 12 Open File Format then you have to have ODF present to counteract it.
Now in a world where Microsoft hadn’t actively ensured that you were going to have difficulties if you used ODF, it would be much more ideal to use ODF – the design point’s slightly different between ODF and Open XML.
Open XML is intended to be a serialisation of the data structures inside Microsoft Office, ODF is intended to be a base line that is extended with functionality so that core documents are then transportable and permanently available.
So I actually think, that ODF embodies the principles for which Government Archivists will use an open format.
So in an idea world I’d use ODF, in the pragmatic world of the Danish market and of Danish Politics unfortunately I think you have to have both.
7:11 – Question
One could actually argue that also from a very pragmatical standpoint there aren’t any problems.
I mean the sort of rationality could be that, since Sun has taken on developing a plug that will enable you to save in ODF format from Microsoft Office and Microsoft is actually working with Novel who is trying to make a plug in that will enable you to save in Open XML format from Open Office what’s the problem?
7:50 – Answer
Well the problem is that neither of those mechanisms actually works for ‘round tripping’.
The point of both formats is that they’re there as a convenience to allow you to bridge between the two worlds.
I mentioned earlier that Microsoft talks extensively about Interoperability when what they really mean is they have created a funnel that gets you into their world.
Well, both of these 'plug ins' are really there to assist you in migration, not there to allow you to engage in a daily dialogue with the other world.
Actually in a previous job I developed a word processor back in the eighties, and one of the things I discovered in the process of doing that, is that the file format of a word processor is an expression of its ‘World View’ and you will always find that there are things in a file format that cannot be expressed in another word processor.
My classic example of this was back in the days of ‘WordPerfect’ and ‘WordStar’ and in ‘WordStar’ ,.. Your listeners may not even remember that it ever existed!..but WordStar was the dominant word processor many years ago and you could centre a word in a line in a document by entering a keystroke and wherever that keystroke appeared the line it was in would be centred.
When we all transitioned to WordPerfect, WordPerfect’s document model was a paragraph based document model rather than a text stream based document model, and when you came to translate that ‘centre this line character’ WordPerfect was unable to actually express it, because there was no concept of a centred line in WordPerfect and it was impossible to implement one, because it broke the paragraph based formatting model.
And so it was actually impossible for that concept to be faithfully reproduced in WordPerfect – so you had to throw it away.
And then when you took the document back from WordPerfect to WordStar, the centred line had gone because there was no way for it to have been expressed in WordPerfect.
That’s a very simple example but those sort of differences in ‘World View’ and ‘model’ exist between all word processors and it is impossible to have a faithful reproduction of the formatting of a Microsoft Office document in any other word processer.
Just as it is actually very difficult, it’s not impossible to have the faithful reproduction of the format of say an Abi word or a Google docs and spreadsheet document that’s been produced in ODF because the design point of ODF was actually to produce a base line document format that would work everywhere, it was not to find a way of encoding every document format that had ever existed.
It was a way of saying that all modern documents contain these features and here is a way of representing those features.
If Microsoft had been responsible, what they would have done would have been to have taken ODF and then using XML namespaces to have layered new functionality on the top, optionally, so that it was still possible to use the baseline format but their added value was then layered on top in namespaces.
If they had done that they would have been actually probably able to gain greater market share and also actually be able to gain respect from the world for doing the right thing.
As it is we are in a world where they have a format which truthfully is not going to be able to round trip anywhere else, so if you look at the plug in that Sun had produced that supports ODF format in Microsoft Office it’s actually very good.
It does a great job of allowing you, it gives you a true ‘save as ODF function’ in Microsoft Word, but on Microsoft Word it plugs in, it adds ODF to the file format, it lets you have ODF as your default file format if you want to, but you shouldn’t expect to be able to repeatedly round trip documents backwards and forwards using that feature.
Similarly Novel’s plug in really doesn’t work very well and you shouldn’t expect to be able to ‘round trip’ with that either.
So those are crutches to use, those are tools to help you migrate; those are not regular freeways to drive your car to work every morning.
11:54 – Question
So one final question, how do you see the present struggle? The present battle over Open standards – where will it end?
12:06 – Answer
Where will that end?
That struggle has been a constant struggle since the eighties.
You know, I was involved in OSI Standards in the eighties and in x400 and I saw how that worked out.
The truth is that whenever you create a system, with that system you create the game that plays it and if we formalise rules for how the market is going to operate unscrupulous players like Microsoft will always find ways to exploit them to their own advantage unfairly.
And the place that this is going to has to be a constant focus and a constant attentiveness to maintaining the competitiveness of the market.
Standards alone these days do not guarantee your freedom, actually, sadly even an open source license these days do not guarantee your freedom.
Right now the state of the art freedom for end users is Open Source implementations of open standards and I would strongly recommend to the Danish people in looking at this debate that you insist that what you standardise on is an open standard that is multiply implemented in Open Source, however else it’s implemented, if there is no Open Source implementation which truly works you should free from it, because the state of the art in lock in is to talk about Interoperability in standards but to have no Open Source and the freedom point for European countries in this particular debate lies at the convergence of open source and open standards.
