> Access to standards also poses challenges. Since the consensus-building process is costly, organizations like ANSI often try to cover expenses by selling or licensing access to standards documents. For example, the book of standards for concrete blocks and masonry structures costs $150. This business model strikes many critics as unjust, since private standards can be built into regulations, yet sometimes only citizens who pay can look at them.
This. Incorporation by reference is a real problem.
Carl Malamud--the lawyer behind law.resource.org and public.resource.org--has spoken a bit on this topic.
Just wanted to highlight this blurb[1] from a PHMSA committee discussing the impact of commercial standards incorporated by reference as it applies to federal regulation:
The only common thread that I see in associations in terms of their business model--and they are businesses despite being not-for-profit, and 501(c)(3)s or (c)(6)s--that you charge somebody to be a member, you use those members and their intellectual property to develop products, and then you sell them back to them. It's a crazy business model in some way. You can't imagine how that would work.
Coding to an API spec is one of the most freeing things I have done as a backend engineer. We use JSONAPI at work, and while it still leaves things open to interpretation and places to fill in gaps, having a standard just frames everything for the better. Tooling can be created on either side with explicit expectations, and we can give new employees a link to the spec to know how things are supposed to work. It's great.
I grew up referencing RFCs in network software land. Then I got a few gigs with eletronics and the first years were the most depressive of my life. every step started with a research for some standard, then the request to purchase it, then in some cases the mailing of some agreement contract, then the wait, to finally see the standard just referenced some other standard, so back to step one.
It's the same in information technology too when you move beyond the actual protocols and languages (which tend to be free and open). You get the joy of having to buy some healthcare related standard from your national registry, only to find that it refers to the corresponding ISO standards with hardly a change, so you get to shell out some Swiss Francs for one of those.
Sensibly enough, the Dutch government made a set of related healthcare IT standards free of charge when the legislature started mandating their use (NEN 7510, 7512, 7513), but they still reference loads of non-free ISO standards for the most trivial of things. You don't really know whether a reference is relevant or not without looking it up.
When you read these things it is always a relief to find a reference to an IETF RFC; that at least means you get a searchable HTML document available at a reliable location for anyone to read.
A pitfall not mentioned in the article: a private-industry member of the standard-setting body may fail to disclose that they are applying for a patent on an essential component of the standard.
For instance, in the 1990s, the oil company Unocal managed to insert a patented technology into the mandatory Reformulated Gasoline standards in California [1]. Eventually Unocal was hit by a class-action lawsuit and an FTC action, but that took years, and in the meantime all refiners in California had to pay royalties to Unocal in order to comply with the state mandate.
This has happened with the ETSI DMR spec. Several key technologies needed to implement it require patent licences from Motorola. On the plus side, ETSI requires patent disclosures, and those are on their website.
A cartel is bad, particularly as long lived as that one.
Then again, if you ever used commercial tungsten bulbs with longer life, they were quite a lot dimmer. Maybe 1k hours wasn't that bad a trade-off. Best would have been both choices in shops.
Thanks for the interesting read. I think the WSJ should have touch upon “open standards” vs. the standards aka patents. They may have alluded to it, but only slightly. WSJ kinda missed the boat here.
Overall I loved this article, in particular the anecdote about construction materials. One sentence jumped out for me.
> These facts should prompt some reflection about the exercise of power in a technological society: Amid concerns about the excesses of market power and government regulation, nobody seems to worry much about the private groups of experts who created 80 percent of the laptop’s standards.
While I think this is overlooked, sure, I don't think this holds very true. I and other engineers I know do worry about which private companies and interests are controlling standards, and what perverse incentives are associated with them.
Recently, there was a big scandal and worry around the NSA creating subvertible weaknesses in technologies deployed worldwide by working on standards committees (a reason academics and the open source community has worked hard to produce alternatives).
Another example that comes to mind are the stardards around Javascript/ECMAScript. During the first few "rounds", there was an anticompetitive effort being waged against browser giants like Microsoft, who were thought to have been pursuing standards choices that favored their market positions.
I'd say there's lower public visibility into these processes, but it isn't like these conversations aren't happening.
> Our modern existence depends on things we can take for granted. Cars run on gas from any gas station, the plugs for electrical devices fit into any socket
As a world traveler I can attest to how untrue that statement is.
Something I liked about the IETF, no cost to get the standard, there was always a running code example. It changed as it go too dangerous.
There isn't any reason to have these big organizations as "standards bodies", all you really need is trademark protection and a test suite. And not even that some times.
I often point people to the Arduino "standard" which immortalized some poor sods PCB layout mistake in spacing the shield connectors.
Basically you define a standard (no consensus required, it just has to be useful). Give it a trademarked name. Offer to let anyone use it on their gear for a one time license fee. Then market/sell/advocate things with the trademarked name until you get people associating the name with the standard thing you have. And at that point you're done. If your licensing price is reasonable (say under $25,000), people will just license it and use it rather than try to compete with it.
The problem is that we are then stuck with whatever design happens to catch on. Standards stick around for a while. All else being equal, it is better if they were designed with some thought to being a standard, and not simply a one-off design.
The last minute envelope (or napkin) design job Pike and Thompson did on UTF-8
Its 220 volts right? not 240, or 250. Or that abominable 110. Anyway, at least its 50 hz. Sorry, sorry, 60hz. Except when its DC and its 48v. Except when the DC is PoE and its 32V to 60V depending...
This. Incorporation by reference is a real problem.
Carl Malamud--the lawyer behind law.resource.org and public.resource.org--has spoken a bit on this topic.
Just wanted to highlight this blurb[1] from a PHMSA committee discussing the impact of commercial standards incorporated by reference as it applies to federal regulation:
The only common thread that I see in associations in terms of their business model--and they are businesses despite being not-for-profit, and 501(c)(3)s or (c)(6)s--that you charge somebody to be a member, you use those members and their intellectual property to develop products, and then you sell them back to them. It's a crazy business model in some way. You can't imagine how that would work.
[1] https://youtu.be/Sdm698P2AkA?t=88