I’m not sure I should care that OOXML got past ISO. If there is a list of things I dislike in the computer world, XML and productivity applications are on the top of that list. I really just don’t use either when I can help it, with the exception of XMLRPC which I grant a pass because it’s very light on the brackets-to-content ratio. Lots of standards bodies have thrown out some very complicated and incomplete standards that really don’t make a developers life any easier. No offense to SOAP, CIM, and WSMAN fans but I’m looking at you here. Maybe you meant well, but the phrase “a camel is a horse designed by committee” comes to mind. Good job on HTTP though, I like that one — and thanks for my intrarweb. If OOXML is now a “standard”, that doesn’t make me any more likely to use it — just for the same kinds of reasons I would shun other complicated standards that don’t help me accomplish my goals.
Still, it bothers me that yet another “standard” for once is accepted that has nothing to do with interoperability or vendor neutrality. AutoSpaceLikeWord95 is a crowning example of this, but all the binary addendums are worse. I think the use of ECMA and Fast Track was equally questionable. Perhaps we are using the wrong way of defining what a “standard” actually is.
What Microsoft is after here is the ability to say “we are standards compliant”, or “we use open standards” — even when the spirit of that term is completely foreign to them — they want to latch on to a bit of Linux’s “openness” for marketing points without actually doing it in spirit. This is the whole “embrace and extinguish” / FUD / misinformation battle tactic. As is the case with many things, it’s easy to pretend to be something, but it’s much harder to really be it. In this case, the word is “open”. I think the transparency of their intentions are still quite obvious and it’s unfortunate that politics drives the decision making in organizations like this. But, in the end though, does it matter what is blessed as a “standard” and what isn’t?
There are a lot of standards that are absolute junk. Even ISO has one. I like their film speeds and shipping container sizes, but let’s look at ISO 9001 for instance. I once had some nice ISO training at a certain large company. Basically it was a 1 hour class on how to respond to ISO 9001 audits. The only line you are supposed to say is “We have a process, we follow the process, and if I have questions about the process I ask Timmy”. What value does that actually achieve as a standard? I don’t know what happens if you don’t have a Timmy (name changed to protect the innocent).
So, I guess the ultimate question is, do standards bodies really make standards? I don’t think they do. I think proven use and success of a protocol/format makes a standard — and that’s the only way. Occasionally you need documentation on how that thing works, but you don’t need a standards body. Look at how successful XMLRPC is — it’s a specification, and widely supported in numerous languages, and pretty much a de facto standard — but is it “formal”? Not really. Does it have to be? Certainly not. Have they wasted time trying to make something official? Not really either. It’s just cool tech that works and lots of people use because it works.
De facto standards — things that work and actually archive something useful are the only standards I personally care about. Go ahead and enjoy your bullet point, Microsoft. If anything good has come out of this debate, it’s that your document standards /still/ aren’t open, and that there are good alternatives. Interoperability, completeness, and neutrality is what matters and OOXML does not provide it. Yes, we /do/ need good and open specifications, there is no contest about that…but the term “standard” was been co-opted.
My point — We should redefine what “Open Standard” means in terms of the spirit of the word, not in terms of organizing bodies. We should use tools that are open and promote transparency and interoperability, rather than reinforce a vendor or market. Think about the end goal of what you want to achieve (avoiding lock in, allowing yourself to change the choices you make later, choosing tech that is at exactly the right level of complexity and no more), not the bullet points on product web pages.
I’ll go back to using vim now 
(Commence flamewar if you want — this blog is moderated so it may take a while to approve stuff)
I should add axiom #73: “Things that have Open In Their Names Often Aren’t”.
According to one of slashdot poster, MSXML approval by ISO is a pyrrhic victory which means Microsoft will have to adhere to this new standard. In long term, how many companies, governments and organizations will be able to implement the Ecma 276 monstrosity increasing the cost and requiring an intensive process? Microsoft proved they cannot even properly support their own formats as evidence with their Office suite.
The ISO organization has greatly tarnished its reputation by allowing itself to be manipulated by a giant company that caused the change of rule. The fast process has proven to be a complete fallacy. Expect an intensive backlash soon.
One of ways make Microsoft suffer is to not use that OOXML format.
Yeah, I’m not sure their own implementation will be consistent… given the purpose of Microsoft’s effort is to be able to claim “standards based” for governments that require it, we can only hope those same governments will fully test interoperability and find no valid examples. They shouldn’t stop at what the product claims it is.
+1
I completely agree!!
byez
Luca
What is even scarier is that they made a total mockery out of the ISO itself. This really implies that the ISO can’t be ‘trusted’ fully the way one might expect.
Fortunately there are organizations such as the OSI that can generally be trusted. Perhaps they need to extend their approach to document formats?
Indeed. With something as notable as ISO falling (I mean “falling”, not failing, but that too), /any/ “official” standards body is now highly suspect.
I am increasingly coming to the belief that standards bodies are useless in software and that only specifications — whether held by groups of people or even benevolent dictators — are actually worth anything.
Do I care about engine screw sizes and the proper widths for railroad track? Totally. Software? Not so much.
Perhaps it is analogous to why in the US we don’t require a PE for Software Engineers — it’s not really Engineering, but just Computer Science / Design / Math. There’s nothing wrong with any of those things, but the traditional rigours of “standards” don’t apply. That should be ok.
Some of the other standards bodies seem pretty bad too — creating standards that often can’t be implemented or just serve to declare an existing single vendor’s implementation a standard. That practice is pretty rampant.
Kinda sad.
Again though, given success of things like XMLRPC (and lots of things like it), perhaps all these things can be totally ignored now. Imagine all the time that’ll free up.
Michael, thank you for your developer vision. In this case, “standards” really don’t matter, only good specifications.
But unfortunately “ISO standards” are what governments tend to use. This is the words they put in their Request For Proposals when they are going to buy things. These nasty “ISO standards” are the words they use to claim how they’ll interoperate and trade across borders. And also how public institutions will interoperate with private institutions. Over time, it defines also how private institutions use technology.
So yes, ISO standards does not matter from a developer standpoint, but they are crucial in shaping the non-developer flow of information.
You, as a developer, can use whatever you want or like. But if you want to interoperate with non-developer folks, you – the smarter guy in this context – will have to use what they use, and they use what have a stronger marketing force as “this is an ISO standard”. This is why good ISO standards are so important.
Yeah, good ISO standards are important, but when the body is compromised *all* ISO standards relating to software lose value.
I still think it’s great that we have standards for car parts and everything — but in terms of software most “standards” are a complete joke — especially unimplementable ones, which seem to be getting to be more and more acceptable.
I am not sure what you mean about non-developers caring. Naturally you do have the government mandate to use “open standards”, though in this case this is just an illusion. If I am not an auto-maker (or working on my car), the idea that the car is compliant in terms of it’s bolt size is immediately less interesting.
Not to be confused with say, certifications and testing, which say it can withstand a crash test. That’s a little bit different — and concentrates on the important part — results — not implementations.
Michael, non-developers are people like my mother, sister, my boss, the sales guy from the other department, the journalist I want to send an article for review, etc.
They don’t even know what standards are. Their everyday standards are defined by the tools they use, and the popularity of these tools are leveraged by marketing statements as “this stuff follows ISO standards”, whatever this really means.
About ISO being compromised, we, the tech guys, know that. But people writing RFPs don’t. They just know there is one single global standardization body and they have been told that standards are good, so let’s just use their standards. They don’t use to think too much about these things…
I would disagree that the proverbial mom/sister even care about standards. The popularity of MS Office is based on “This lets me edit documents”, and somehow, people pay them lots of money for that. I don’t think it will be any more/less popular with those people because of an ISO standard.
Anyhow, yes, it’s important for RFP marketing bullet point compliance — for the governments that now mandate “open standards”. So, then, as you said — it’s just a bullet in a marketing presentation — what’s the value in worrying about how good a standard is at all?
That was my point.