[Random-bits] Is Open XML a one way specification for most people?

[James Love]

http://www.sutor.com/newsite/blog-open/?p=3D1145
Bob Sutor=92s Open Blog
Monday, October 16th, 2006 @ 11:56 am
Is Open XML a one way specification for most people?

I have been accused in the past of using a "weight" argument against =20
the Open XML specification because it is several thousand pages long. =20=

While some people may think that is cute or funny, it is a real =20
concern and is an obvious problem that programmers recognize. That =20
is, it is hard enough to implement a standard correctly, and one that =20=

is that long will be virtually impossible to do right. Mark my words =20
on this and let the buyer beware.

Who will implement Open XML correctly and fully? Maybe Microsoft. =20
Why? Since it is essentially a dump into XML of all the data needed =20
for all the functionality of their Office products and since those =20
products are proprietary, only they will understand any nuances that =20
go beyond the spec. The spec may illuminate some of the mistakes that =20=

have been made and are now being written into a so called standard =20
for all to have to implement, but I=92m guessing there might be a few =20=

other shades of meaning that will not be clear.

Fully and correctly implementing Open XML will require the cloning of =20=

a large portion of Microsoft=92s product. Best of luck doing that, =20
especially since they have over a decade head start. Also, since they =20=

have avoided using industry standards like SVG and MathML, you=92ll =20
have to reimplement Microsoft=92s flavor of many things. You had better =20=

start now. So therefore I conclude that while Microsoft may end up =20
supporting most of Open XML (and we=92ll have to see the final products =20=

to see how much and how correctly), other products will likely only =20
end up supporting a subset.

That means that other products and software, in practice, will NOT be =20=

able to understand arbitrary Open XML that might be thrown at them. =20
There is just too much. Therefore they will only create a bit that =20
they need and send that off. Send it off to whom? The only software =20
that might understand it, namely Microsoft Office.

So this is how I see this playing out: Open XML will be nearly fully =20
read and written by Microsoft products, but only written in subset =20
form by other software. This means that data in Open XML form will be =20=

largely sucked into the Microsoft ecosystem but very little will =20
escape for full and practical use elsewhere.

In my opinion, suggesting "choice" among ODF and Open XML by =20
governments who are seeking control, true choice, and =20
interoperability is nothing more than maintaining the status quo =97 a =20=

requirement for Microsoft products under the guise of supporting a =20
"standard."

All standards are not the same and providing support for all =20
standards is not the same. Think about the implications of what is =20
going on here. Open XML is not about interoperability in general, it =20
is about moving or keeping a vendor=92s products at the center of a =20
universe. Nice marketing tactic. And that=92s what this is about: we =20
hear nice words about "open" and "xml" and "standard" but the reality =20=

and problems of producing real implementations are left to technical =20
folk. Listen to what your technical people tell you before you make =20
policy decisions that are not open and not particularly practical.

To be clear, people can choose to implement Open XML or not. I think =20
some will try. Let me know after you do.

Finally, do this little thought experiment: imagine how thick a ream, =20=

or 500 sheets of paper is. Double that to get the thickness of a =20
thousand pages, make that 4 times thicker to see how thick 4000 pages =20=

is. That=92s how many pages were in the last draft of the Open XML =20
spec. How many people will you need to implement that fully and =20
correctly, much less read it? I believe the final version is around =20
six thousand pages (correction?). I think we=92re already past =20
feasibility for most people unless you=92ve already implemented and =20
debugged the software over a period of years.

---------------------------------
James Love, CPTech / www.cptech.org / mailto:james.love@cptech.org / =20
tel. +1.202.332.2670 / mobile +1.202.361.3040

"If everyone thinks the same: No one thinks."  Bill Walton"