What is a standard?

A question that comes up in different forms in my work is “what is a standard?” The answer is not as straightforward as one might imagine.

My definition is that a standard is the consensus of the people who wrote and use it; where ‘consensus’ is the highest common denominator.

If that’s what a standard is then I should perhaps give an example of something that isn’t a standard. Step forward project teams that publish their finished work and say “here’s our standard, you can use it.” That’s not a standard, that’s a mixture of documentation and wishful thinking.

There is a gradation between the two and it’s not always easy to draw a distinction. The UK Government made a good attempt at it in 2013 and updated its definition in 2018.

A standards-like document written in public (usually on GitHub) to which anyone can contribute, and that has a lot of support, is clearly close to my definition of a standard. However, there are two crucial aspects missing: 1) a formal process; b) intellectual property guarantees. The formal process is a user's guarantee that the standard is stable and can be built upon. Change can only come through following that formal process and it's that that also provides the assurance that implementing the standard will not result in unexpected lawsuits. Occasionally, a standard will emerge that was effectively written by a handful of individuals, without any formal process, but that becomes so widely adopted that to say that it's not a standard would be silly. RSS is an example of such.

Note that the UK government definition refers explicitly to open standards. Can a standard be closed? Well, yes it can. Your mobile phone, for example, is built on a large number of standards but their use is licensed at a cost, so they’re not royalty-free and therefore not open according to the UK government definition. Open doesn’t mean that the standard has to be free to access. You have to pay to download a PDF from ISO, you don’t have to pay to read a standard from SDOs like W3C, GS1 or OGC. But, there’s a quid pro quo. You need to be a paying member of those organisations to be a voting member of a working group – there is no such fee at ISO. You either pay to write or you pay to read.

Does that mean that standards development bodies are in it for themselves, developing standards for the sake of generating membership fees on which their salaries depend?

I am very confident that the answer is no - if only because it simply wouldn't work.

On day 1, membership organisations exist to serve their members. From day two, they serve themselves.

I heard this quoted by Esther Rantzen when she was being interviewed on Radio 4. I am very sorry to say that I can neither remember the programme nor, more importantly, the name of the university professor she was quoting - I'd love to know. I can confirm that it wasn't during her March 2019 appearance on Desert Island Discs.

Having worked for a succession of membership organisations this struck a chord and I can confirm that it is true. The implication of the quote, however, is that all membership organisations 'go bad' as they serve themseleves without serving their members. That is not true in my experience. That is, serving a well-run membership organisation does not go against the interest of the members. For every time one hears "our mission is to make ourselves redundant" you'll hear "if we didn't exist, our members would have to reinvent us."

Working for a standards body means being sought out occasionally by people wishing for their prized work to become a standard. The answer is always “which of our members wants this?” It's only if the answer is “lots of them” that the standard will be put through the system and during that process it will change. That's the SDO acting on behalf of its members to run the process and create a stable, no-IP surprises, foundation on which they - and you - can build. If that's not true, it ain't a standard.