Wednesday, August 19, 2009

Developing Standards with an Open Source Model - Possible? Any Interest?

Rob England (@theitskeptic) retweeted (and agreed with) a Tweet from Antonio Valle (@avallesalas) who said "Why not create an open-source alternative to ITIL?" I responded asking, "What about an entire open source standards movement?"

I have, in the past, been involved with both ISO and IEEE software/systems related standards work. Without getting into an entire explanation of how standards bodies work, one of the issues raised by people on the "outside" is that it is very costly to get standards and difficult to find out the real status of ongoing work. At a certain stage in standards work, drafts are no longer available to anyone outside the standards working groups and voting bodies.

To be involved in such groups costs money, not because there is a high membership fee. There are 2-3 meetings of each per year, which means travel & living costs plus time away from work for 2-3 days each time. This cost dwarfs any fees associated with membership. Consequently, membership in such organizations usually means corporate support of some sort. Therefore, most of the people involved represent corporate interests. In the case of the software/systems areas in which I was involved (process related), most participants were either government bodies or contractors to the government (or consultant groups associated with government work).

The argument for the cost of standards is that the bodies overseeing the work have legal and administrative costs to cover (besides publishing ones) to ensure the fairness of the standards work. This fairness is very important since contracts and business often depend on meeting standards, so making sure the standards do not unnecessarily favor certain vendors or industry groups is very important. (In the case of ISO standards, the issue is to ensure countries or blocks of countries are not so favored.)

However, standards are made by those who show up either to the write or vote upon them. The former carries a good bit of weight since voters rarely spend the time to re-write what comes out of a working committee, though comments can be extensive. Part of the fairness issue is to make sure all comments are addressed and dealt with appropriately and in an open manner.

It can takes years for a standard to go from a proposal for new work to an actual accepted document. So the cost can be significant to participate, in money and time.

Now one way to have an impact on this process, is to come to the bodies with a draft of a standard already in hand. Of course, that is usually done by some member of the group or some document "sponsored" by such a member. Once a document is created by or adopted by a standards body, the copyright belongs to that body, i.e., they "own" the document. From then on, they are the ones entitled to make changes (or not) to it.

My thought when I asked about an open source standards effort would be to see if a standards document could be produced that could get industry support but be developed in an open source manner. I did some work on an agile-related standard for IEEE where I attempted to use email rather than face-to-face meetings to try to get broad international participation. But it was not open source and consensus was difficult to impossible to achieve that way with the large number of people involved. An open source approach might be able to overcome this as the open source community has addressed software developed in a "community" fashion.

Do people who have had open source and/or standards experience feel this could work? One major question would be how the results are handled and appropriate "stability" introduced so the result can be reliably used as a reference that is not changing too frequently given how standards are used over many years and may become part of contractual considerations.

No comments:

Post a Comment