I will not pretend that M$ honestly wants to be “open”, and just can’t believe there is people who believe it’s going to be. Is it a good thing that M$ opened its specification? Yes, it is, of course. But please don’t blindly believe that itself alone makes it necessarily good without considering the motivations of the company.

Thanks Rob, you have so great damn good arguments!

We should also remember how hard Microsoft was pushing the Office 2003 Reference Schemas in Massachusetts even though the spreadsheet formulas were totally unspecified.

Of course, hypocrisy is not a crime.

“Also some commenter here suggest that the OOXML spreadsheer formula’s were based on the work of the OpenFormula’s group. However he seems to completly overlook that fact that the OOXML spreadsheet formula’s were based on SpreadsheetML, a predesssor of OOXML which was created in 2001/2002, even before the conpcept of OpenFormula ever existed.”
# posted by Anonymous

No matter when SpreadsheetML was released, the fact remains that the XML encoding of Excel formulas were not included.

It’s an interesting fact of history that Microsoft did not disclose their Excel fomulas until after IBM had contributed the complete Lotus 123 formula and spreadsheet conversion documentation to the ODF OpenFormula Project.

This is hardly what one would characterize as the straw that broke the camel’s back. It’s more like a ton of bricks falling on Microsoft’s head. They had little choice but to finally give into public outrage and release the secret formulas.

Note that if Microsoft truly believed in the application interoperability they now preach, they could have released the Excel formula documentation at any time – including the 2003 MSOffice release of SpreadsheetML. But they didn’t.

OBTW, now that the MS Translator plugin has been released, suspiciously coinciding with the final days of the Ecma 376 contradiction review, does anyone think Novell will release the Translator plugin for OpenOffice.org on Monday, the final day of MS Ecma 376 review – when fast track ballots must be cast?

~ge~

But if you think it makes a bit of difference, I joined IBM as part of their purchase of Lotus, and over the past 17 years I’ve coded on SmartSuite, Lotus Components, eSuite, K-Station, Portal and Workplace, among some of the more notable projects.

Along the way I’ve done a good deal of file format work, with SmartSuite formats, with the legacy binary Office formats, with ODF and with OOXML. I also was part of the team that made the Xalan XSLT engine and contributed that to Apache. So I have some basis for my opinions, based on practical experience as well as clear thinking. You are free to accept or reject either or both.

To the point of interoperability, it is a more complicated issue than you make it out to be. The presence of a specification does not guarantee it, and the absence of a specification does not prevent it. It is a matter of degree based on the transparency of the technology. A binary format, for example, is very difficult to reverse engineer. But a spreadsheet function is rather trivial to do so. The concept, familiar in patent law for, of “undue experimentation” is appropriate here. Can you figure out what AVERAGE() means without undue experimentation? I think so.

Also you might want to take a look at the OOXML formula documentation to see if it does much better. It doesn’t tell whether SIN or COS take radians or degrees as input parameters. So obviously even there interoperability will require the use of information that is beyond the normative text of the specification.n.

OOXML (Nonsense and useless)…we have ODF which has been developed by multiple vendors WHICH IS OPEN and ready for everyone to start using…Why do you insist that we should work with OOXML. ODF is here to end this madness?! OOXML is there ONLY because ODF was gathering momentum…its a bit like IE7 only came around because Firefox has started to take market share?

The real problem that I see is that you are going to need to support XAML with Mono (at some point!) explain how you are going to do that and not touch on any patents? I mean…I keep reading that Mono is not patent encumbered (are you sure about this??) I can only say that it is your lack of wisdom when it comes to dealing with Microsoft…you need to be stung in the a$$ so to speak to realize who you are holding hands with?!?

I complained about it not only to Brian Jones, but also to Port 25. I was surprised to see that they had taken my complaints on board; I had suggested either they provide the full specifications for ActiveX, or abstract the functionality so that I could get the same sort of effect using QT or GTK. They chose the second option.

Now all the hard work by the Groklawyers is further proof that Microsoft hasn’t put nearly as much thought into their Office Open XML file format as they fondly hoped.

My experience is proof that Microsoft suffers from a Not Invented Here syndrome, and while it is possible to get them to change, it takes a lot of work.

Microsoft has already launched their MS Office 2k7 product, with MS OO XML aka ECMA 376 as one of its default file formats. And there’s a whole lot of other equally serious issues with it, that Microsoft shows little or no interest in rectifying.

Which is why, when I get back on to my (stalled) Office Miniatures project, I’ll be using ODF as the default file format. I can always read the KOffice and OpenOffice.org code to see alternative ways of handling major problems.

If anything, his experience with Gnumeric should have taught him that interoperating with ms products is a lose-lose situation. A constant up-hill battle of catch-up which merely burns through engineers without ever quite being good enough to compete. The 6 000 pages of ECMA 376 should clearly demonstrate this continues to be a unachievable goal.

MSOOXML is clearly indefensible as an open document interchange standard on any technical ground, and it is dissapointing that he is taking this highly political position. Coming as it does at a time when there is a limited, yet real opportunity to take advantage of the current momentum in the free software and open standards realm, and the unattractive cost (and retraining) of moving to vista and office 2007 – it is also irresponible given his position as a free software personality.

There seems to be some confusion with regards to file formats and applications. Either I’m confused, or others writting about the subject are confused.

Simple non-confused statement: ODF is a file format specification.

The confusion seems to be around building applications against the specification as opposed to using the specification.

For example:
Prove you can build a *COMPLETELY* compatible Excel/ooxml program based just on the ooxml “specs”.

A file format specification identifies such things as:
a) What bits are user data.
b) What bits are meta-data about user data. In other words, is that value a name, or a date, how about a birthdate.
c) What bits are meta-data concerning the software itself. As an example, you may have such things as how the user-data is to be formatted.

One could build a straight program that uses it’s own file format. Or that same program could be made to read only the user-data from that particular format.

I’m a developer. I write code to handle reading and writing to files, along with their formats, all the time. The most common format I use is flat ascii text. To the non-programmer, an example:

User-Data
Name: John Doe
Birthdate: 01-Jan-2000
Name: Jack Smith
Birthdate: 01-Feb-2000

Flat ascii format
John Doe 20000101
Jack Smith 20000201

The description of that file layout would be:
Name Birthdate
xxxxxxxxxxx yyyymmdd

The above shows that the list of x’s is how long the name can be along with what value fits that “field”. The second “field” is identified as the birthdate and it should be in the format of a 4 digit year followed by a numeric 2 digit month follwed by a 2 digit day.

With the above example, what my program decides to do with the user information stored in that file, the file itself doesn’t care.

Let’s say I wanted to produce a mailing list, I’d want the user-information that defines the individuals along with their addresses. My program, using the file specfication, knows where to find that information.

Let’s say I wanted to tally up any outstanding debts. My program, again using the file specification, knows where to find the financial information.

A file format is NOT the application/program. An application/program should NOT be used to define a file specification.

Now, any developer that has to send information from their system to a third party where the two have agreed on a file layout is very familiar with what I’ve outline above.

So…. Either I’m really really really confused about what file formats are supposed to be for, or MS has succeeded in convincing people that a file format specification describes the application. If MS actually follows that practice, that’s a very bad practice to have.

RAS

But is the Office for Mac team working from the OOXML spec, or are they reverse-engineering or looking at code from the windows version?

I would also note that Office 2007 has only been around in release form since 11/30/06, so my thoughts about it and OOXML having dissimilar behaviors and definitions should not be dismissed.

Has anyone audited Office 2007 against the OOXML draft? Fed valid OOXML to Office 2007 to see if it crashed? Or checked the output files to see if it resembles what is specified on OOXML?

When OpenOffice and KOffice did something different, it was either a bug in one or the other or far more importantly, an ambiguity in the specification. (and I suspect the spreadsheet formula handling is also being done with a Null against reference approach). Without a second implementation, the ambiguities stay just that, fast-tracked into approval. And unlike legal contracts, ambiguities are not decided against the author.

(I was involved in RFC2440 and wrote a very lightweight but 100% comprehensive implementation and regression test and found a small bug in one of the others – the lesson is more implementations, fewer problems).

Microsoft could document every behavior of Internet Explorer in detail, but that would not make it W3C compliant. But they could take that list and submit it to ECMA and ISO as a competing standard for web page authoring. Don’t we need another? And it would be open too! IECSS2 and IEHTML anyone?

So is Internet Explorer 7 fixed so that it properly renders a page which passes all the validation tests and looks right in Firefox, Opera and the rest?

Those standards have been around longer than OOXML. And they are clearer. And isn’t Microsoft on the committees?

How is MSN coming along?

I suppose it will be interesting to see if Microsoft can follow a standard of their own writing better than one they just sit and watch or seem to purposely ignore.

And then it would get interesting if a second implementation (preferably GPL) did something correctly but Microsoft refused to fix Office 2007+ to obey their own specification (and if it broke documents). Would governments and others then be forced to switch to such OOGPL?

I don’t have Office 2007, or I’d start looking now. But either way I think it would be interesting – Microsoft correctly documented something or they aren’t OOXML compatible.

I think you need some reading comprehension lessons. I looked up this blog:
http://www.oreillynet.com/xml/blog/2007/01/what_is_contradiction_of_an_is.html

Here’s his conclusion:
“(I should stress that even if you do accept that “contradiction” has the kind of very specific meaning that I describe here, and even if under this meaning the supposed contradictions detailed by Groklaw fall down on inspection, and even if it means that OOXML can pass beyond the Feb5 adminstration periood into the next phase unmolested, it says nothing about whether OOXL should be defeated (or accepted) at the final ISO vote by member bodies.”

If you bother to actually read his comments and compare the original list, you’ll find that a number of things on there (like, oh, the leap-year problem) DO completely match his strict definition of contradiction.

Microsoft has plenty of money. You’d think they’d be able to hire better astroturfers.

Dude, with all that experience and wonderful OOXML documentation why don’t you write ooxmlmeric tonight? Prove you can build a *COMPLETELY* compatible Excel/ooxml program based just on the ooxml “specs”. FULLY COMPATIBLE or don’t bother. Put it out as open source under GPL and offer to fully indemnify anybody who uses your program or touches the code or uses the code in another project, without any financial limit and including all their legal fees. Your employer is big on indemnity – ask them to do it. Then your comments will have some credibility. Right now, with the many compatible implementations of ODF spreadsheets (Huh? with that flimsy spec that you refer to?), you are just not credible. Here’s your chance to show everybody wrong TECHNICALLY AND BUSINESS-WISE.

Are you up to the challenge? Or are you an armchair shill?

As for OOXML, even if it provides us with all kinds of information–even information that will help us interoperate with Microsoft–that doesn’t make it a good standard. I see it as an unnecessary duplication which is irrationally tied to one vendor’s internals, while ODF has already grown beyond that.

There’s no need for senseless duplication. OOXML just isn’t ready to be any kind of standard.

We think OOXML would be a disaster as an ISO standard, on many levels, and so political involvement is unavoidable. Indeed, the whole concept of labelling something a “standard” is a purely political and social matter, and has nothing per se to do with how something is technically implemented (except insofar as ISO imposes quality controls). And the consequence of labelling something a “standard” is also political: social impetus for developers to follow one specification as opposed to others. Nor is Miguel’s post any less “political.”

Miguel seems to have fallen for the common propaganda that “political” (i.e., trying to persuade others to cooperate towards common goals) unavoidably means “foolish” or “dishonest” or “at odds with technical goals” or “irrational”. That we should all be silent worker bees (as if choosing to be silent were not also a “political” statement about what one values).

Thanks Rob for your detailed rebuttal, and Miguel: the earlier you jump fences the better for all of us.

R.

You say that you are an open source advocate so why do you think promoting XAML/OOXML (which is getting enough underhanded marketing at the moment) is actually going to improve open source and open standards. In the words of a true english developer…stop talking bollocks!

miguel – I am an armchair general.

when it comes to MY DATA I AM the general. and you people that write the software better realize that before you go out of business.

“There is a good case to be made for OOXML to be further fine-tuned before it becomes an ISO standard. But considering that Office 2007 has shipped, I doubt that any significant changes to the file format would be implemented in the short or medium term.”

thanks for letting us know – seems like you have the inside track on where office is going to go. this sounds like a steve ballmer threat.

good luck with that microsoft affiliation and DRM.

This could be extremely laborious and, in many cases, impossible. Consider the compatibility options in docx files. Many of them, such as WordPerfect 6.x justification, do not represent simple behaviors but complex algorithms. The only way to convey these in ODF would be to extend ODF–and then you have the Microsoft-Java situation all over again, “extend-embrace-extinguish.” Is that what ODF supporters want?

2. “I’d like even more for them to make the specification for the Office binary formats public”

That is not going to happen. Microsoft cannot publish these formats because a) reveal too much about their programs’ inner workings (they’re akin to memory dumps) and b) they are encumbered by all sorts of third party licensing problems.

Besides, do we really want to keep these formats alive? They’re a mess. Let them die out!

I actually *wrote* a spreadsheet and I actually *know* how the formula implementations came to be.

There was a lot of guessing and reverse engineering to get those in place.

My point, which you sophistically tried to ignore is that ODF lack of fundamental pieces of the standard would not allow for an implementation based on the spec without resorting to third parties that have reverse engineered the code.

Other than trumpeting ODF, have you actually contributed *ANY* code to OpenOffice, or you are just another armchair general?

Miguel.”

Appreciate everything you did for open source and I wish you luck in your future but one thing I have realized is open source is bigger than any one programmer and that it will survive even without Linus.

ODF is a simple standard to create products from – as devices get smaller I think this will become even more important – small portable open standards to write products from.

I think ODF is a better standard than microsoft’s. I have read both and if I was going to implement a product from one it would be ODF.

If you have any poll with microsoft I would like for you to suggest to them that they support ODF in their product. With all that money and since it such an inferior, simpler, and shorter standard I am sure they would have no problem implementing it.

“It seems like it would prevent some nice features developed in the last 8 years for no other reason than “there was a standard for it”.”

that is a good enough reason for me. If missing those “innovations” means not locking me into one company’s product then you can have those “innovations” I value choice of products more than fancy graphics.

Here is an example for you – was working on documentation for a solaris upgrade project at work and I did it in ODF exported to a pdf file – had to work on it at home and luckily my daughters Imac has ubuntu linux on it and koffice but not open office because it is somewhat limited in resouces. anyway I was able to work on it through koffice because of ODF. this is what I want from a standard. and being in IT I think this is what we should provide to our customers. Not lock in just for some innovation that they don’t care about and have to pay more money for just so someone can buy another yacht.