While I agree that standards can be a tedious PITA to work, and can be terrible, you’ve phrased it with a ridiculous sweeping breadth that I can’t agree with.

I think your initial assertion that standards are useless crap because they cannot “force someone to implement something correctly” is completely absurd. It’s like saying that an orange sucks because it cannot repair a car. Be realistic about what a standard actually is. If you want to try to force anybody to do anything, then the closest you’ll be able to do with words is a legal contract. Standards (in any form – from ISO/IEEE standards on paper or verbally agreed standards) don’t to force people to do anything. Even a “perfect” (i.e., completely crystal-clear and unambigious and self-consistent and so forth) standard will never succeed at forcing anybody to do anything because it is still *only* a guide to interoperating with ideal applications that follow the standard perfectly, and not a legal document. Naturally in the real world there is clearly no such thing as an ideal version of *anything*. You can complain that the world isn’t perfect, but that won’t change anything.

And then there’s “…in the end, it doesn’t matter if your application is standards compliant, what matters is that it works.” Well, sure, unless your system needs to interoperate with other systems, in which case odds are it’s going to have to do so in a way that the other systems understand, and your system will have to be de-facto compliant even if nobody puts a “compliant” stamp on it. If everybody decides that the standards don’t matter, then interoperability will suffer because the applications will all be speaking different dialects of the same language that other applications don’t expect. And that’s not good, because flaws in the standards and implementations make it difficult enough in the first place, without even further difficulty added when engineers and marketing divisions decide that standards aren’t as important as what they’re trying to accomplish, and introduce incompatible variations.

If a standard doesn’t work, then it ought to be replaced. If engineers start to believe that a “x and y and z standards suck” means the same thing as “standards suck,” then interoperability will suffer.

Chris: Well, if Mozilla’s HTML implementation were the standard and not the W3C doc, you could figure out every problem you had easily: write a test and run it in Mozilla. And if you’d still didn’t get it, step through the code in a debugger. WIth a standard, you can’t run it, you just have plain text. You have to ask the people that wrote it what they meant. And if those people are unresponsive, you lost. So HTML is only as good as the people around it make it.
And for the usefulness of the HTML standard, look at how many people sniff for the standard. They only sniff for browsers that implement something they want and then code to it. If the OOXML standard says #FF0000 is red and MS Office reads #FF0000 as blue, guess what other apps will do.

Not exactly cause for celebration. Great to see Jody working to make the best out of a bad situation though.

– Chris

Standards are massively useful. I’d wager that the vast majority of people who visit freedesktop.org do so as spectators looking to follow standards, not developers wishing to collaborate on them. The difference is between standards and “standards”. Just like it is between open and “open”.

– Chris