Friday, November 7, 2014

Lessons Learned on Standardisation at a Finnish Standards Meeting

A few weeks back, I was invited to a Fisma (Finnish Standards and Measurement Association) meeting to discuss #stop29119. Fisma is the Finnish counterpart for international standardisation, participating as one member country in all the different committees someone ends up being interested about. The meeting was futile attempt to do anything about an impossible theme, but I came out with some lessons I considered interesting.

With what I learned there, my summary is this. The whole standardisation thing is based on fooling customers of standards into believing they might be buying something useful and respectable. Fooling, because the process of creating standards really sucks.

1. Power of individuals who volunteer in creating standards

Listening in for the whole meeting about deciding on standards and introducing ongoing work in the area of that particular working group, I came to realise how weak the standards creation process really is. It seems a standard emerges from someone having an interest to spend time on writing a standard and finding an appropriate (politically acceptable) title for it and then working on the contents through a great number of stage-gates.

There is no requirement that the standards authors would actually be really experts in the topic. It seems quite common that some professor / researcher from a random university volunteers time and effort on creation of a standard, and if people in countries are not really strongly against it, the process flows forward through various votes.

Countries decide which standards they participate on based on people's interests they themselves have available. If there's an agile standard on the way (and there is, scary!), Finland will participate if there is someone volunteering for the work. Finland (or any other country for that matter) opting out from the work does not mean that the work would be rejected. Rejecting requires active work and more often end result of disagreement is to create just one more standard with a slightly different title.  Organisations pay for being in positions to volunteer and organisations pay for the end results to finance a relatively complex system.

There is very little quality built into the process of creating the standards, quality is left for paying users to assess. The requirements of expertise are not exactly high for entry.

2. Early detection of problems only

The standardisation process is a process with many stage-gates. It was interesting to listen to discussion where a comment would be "how did we vote last time, we can't change our mind at this point". If you plan on changing the contents or getting the standard rejected, you have to be voting against it consistently from the beginning and build a case on it not improving. You will also need allies from other countries. It was interesting to hear that "Japan and USA voted against but this still goes forward - they always vote against". And still the standard gets created unless more countries are against with severe observations marked for the review process. With 28 countries voting with the process requiring severe observations to be allowed to vote against, getting a standard through does not seem that complicated. Voting to disapprove you get to go into a hearing - that had very negative connotations in the discussion. And being passive is approval. The process is awful.

Things such as early bad quality hiding more severe problems was not visible at all. If I have to read a document avoiding gaping holes, it is highly likely that the large holes steal away my attention. Regardless, changing one's mind is not encouraged.

3. No drawback mechanism and how a standard dies

#stop29119 calls for withdrawal of the standard, and I learned that the process of standardisation really includes no such mechanism. Standards exist until they become obsolete, if they have been accepted. They become obsolete if during 5 years a new working group to update - even slightly - the standard is not formed.

The only way for a standard to become obsolete is that no one volunteers to work on it. Since volunteers are again representatives of paying organisations, those who have interests in standards existence will drive the updating. A standard seems to only die of no one is willing to contribute into it financially.  #stop29119 people buying the standard for review purposes actually contribute financially to the standard to continue to exist. Not participating, not buying are the ways to kill a standard and even then, someone isolated can quite easily keep it alive as long as there's a business to be made on its existence.

4. Standards are just guidelines, you don't have to use them if you don't declare compliance. 

The whole discussion about that the word "standard" means seems like a bad joke after the meeting. Standards are all optional, so the standardisation process itself does not include the idea of it actually being required for compliance. Compliance comes from a user of a standard requiring compliance downstream. Standards are just prepared paper piles that create work for those who keep them alive. We waste our efforts into thinking that standard would be of any value, and that it would be based on anything other than the idea of someone being fooled into using it with vague marketing talk.

5. Thinking standards show an area is worthy

It was interesting to hear remarks that Testing is now worthy area, as it finally has a standard. That areas of relevance gather more standards. If a standard does not apply, an optional other standard can be created with just tweaking the title for more specific case.

The way standards are constructed is that you don't really mention the context / application area but leave that for the users of the standard to think about. Which again underlines the idea that you must be a fool to pay for a standard and that those who consider using the standard are the real audience to talk about what is bad about a particular standard.

CONCLUSION

So, lesson learned. Don't buy the standard. Don't finance it. Don't comment on it to help improve it. It might wither away and die in 2016 if no one starts a new working group to update it. As the Finnish national board participates annually only in 10-20 % of standards ongoing, we mostly quietly vote for yes.

The change must start with the users. Other than that, we waste our times and efforts on something that is just rotten.