In his essay The Myth of Free Standards: Giving away the farm, Andrew Bank proposes that standards should not be given away, that instead their users should be charged. The reasons given for his making this claim are that free standards will "eliminate the most significant source of funding for standards development [and] create a greater imbalance of support between contributors and non-contributors." He also claims that it will reduce the incentive for private innovation, at least for the niche market of those selling access to commercial standards. Finally, he claims it will undermine the perceived value of standards by lowering the price tag.
Let's get biases out in the open now. At the time of my writing, Mr. Bank worked for CSSinfo, which "is an engineering information service that provides specifications and standards." I, on the other hand, work as an independent consultant for Internet technologies. I interpret the appropriate Internet standards for my clients, helping them leverage those standards to improve their business. Although freely available standards are anathema to Mr. Bank's business, they are the lifeblood of mine.
To begin with the issue of innovation. Mr. Bank claims that free standards stifle innovation in his field. This is true, as if something is freely available, there is no incentive to produce innovations to improve its availability. However, other industries rely on information retrieval as a critical technology, and innovation of that technology will continue in those industries in any case.
The thing to note is that the innovation being stifled is not in the area described by the standards - it's in the area of Mr. Bank's business. This is because developing a standard is not an act of innovation; it is an act of politics and writing. Although that process can be innovative, or the information in the standard can be innovative, the standard itself is not. A standard is, at best, a snapshot of the current state of an art. As such, a standard represents the opposite of innovation - it is static document, not undergoing development.
For innovation to proceed after a standard is published, the innovators need the information in the standard. Without it, they won't be innovating, they'll be reinventing. The costs of those standards can kill projects.
For example, suppose you have an idea for an innovative new web application you would like to develop. Your initial capital outlay includes a computer - costing roughly 500 US Dollars at this time, assuming you don't have one. You also need an appropriate programming tool. For that, GCC, Perl or Python might do - all of which are free. You also need the information in the standards that a web application must obey: HTML for text markup, HTTP to talk to web servers, FTP to talk to older file servers, SMTP to deal with receiving electronic mail, POP and IMAP to deal with sending mail, NNTP to talk to network news servers, and MIME to understand what is being sent over all those channels. Of course, if you want to display graphics, you should have copies of the standards for GIF, JPEG and probably PNG. Since these are standards in an area still undergoing innovation, you will also need access to the standards describing various extensions and extension mechanisms for most of them. Finally, HTML is based on SGML. At this time, all of the standards listed are free except SGML, which costs 236 Swiss Francs, or 133 US Dollars. If the other eleven standards cost the same, including the extensions previously mentioned, that would come to 1,463 US Dollars. We've just raised the cost of starting our nascent business by almost factor of four. Maybe the world doesn't need an innovative new web application just yet.
Although that sum - less than 1,500 US Dollars - is inconsequential to any company that is already a going concern in a developed country, that is not necessarily so for other companies. In less developed countries, this amount represents the production of several people for a year. In any country, a company that consists of two people working in their garage would have to take the money out of a family budget. Since the latter is how such companies as Microsoft, Apple and Hewlett-Packard started, these companies cannot be dismissed out of hand.
Since Mr. Bank's claims that free standards have both funding problems and an unfair distribution of cost, the question of how all the free standards mentioned above - and the thousands of others I didn't mention - are paid for. Many were developed by the working groups of the IETF. These working groups are composed of people who are on the mail list for that working group. Their time is donated to the group because they, or their employer, is interested in the goal of the group. When the standard is complete, it is published by the IETF by making it freely available on the Internet.
Before looking at why a company would allow an employee to spend time on a document that is going to be given away, let's look at who benefits most from a standard. The answer is rather surprising, because those who benefit most from the existence of most standards never directly pay any of the development cost. That's because those who benefit most are the consumers of the products that use the standard. Every time someone purchases a product built with bolts that have standard threading, they benefit from the standard that describes the threading. Every time someone loads a web page into a browser, they benefit from the existence of the many standards that come into play in the process. The consumer is paying indirect costs for those standards, as the overhead of developing - or accessing - the standard is included in the cost of the product, or of the computer software they are using.
In that last sentence is the key to why a company would support the cost of developing a standard that will then be given away. They benefit from the existence of the standard. Without a standard for exchanging electronic mail, there can't be a market for software to exchange electronic mail - either the client on the users desktop, or the server running at the ISP. By funding the development of such a standard, the companies involved create a market for their products.
The question then arises of whether it's fair that companies that didn't help pay for the development of a standard have free access to the information it contains. After all, doesn't this lower their overhead, thus giving them an advantage over the companies that funded the standard? The reality is that they don't, because the playing field isn't level. The companies that helped pay to write the standard got to write the standard. That means they had advance access to the information contained in it, and a chance to have it describe products that they had already developed. In extreme cases, a company can write a standard that describes an existing product of theirs, and submit that description to the IETF.
As a final point, the question of obtaining printed copies of free standards should be addressed. The IETF avoids the cost of printing, publishing and distributing printed standards by not doing any of those things. They anticipate that those wanting a printed copy will create it themselves, using resources that are readily available in even a small office. Those wanting a higher quality edition than can be produced locally provide a business market for independent publishers. Although the cost will be higher than if there were no free competition, it may well cost less than a printed copy of a commercial standard. Witness The SGML Handbook, which includes the entire text of the SGML standard, yet cost only 95 US Dollars - 35 percent less than the standard!
Thus, we have a rational model for funding standards development that is fair, and has been demonstrated to work - and work well. If it didn't, neither my essay nor Mr. Bank's would be on the web, because neither the web nor the internet would exist. Yet Mr. Bank fails to even mention this model for funding standards development.
Mr. Bank quotes the "you get what you pay for" mentality to try to show that a standard being free lowers its value. I note as a counterexample that his own essay suffers from this effect, being freely available. Yet it still has enough value to be worth replying to. None of his supporting examples involve standards, or anything that would be critical to a company's business plan. However, to those wishing to provide products that must conform to a standard, the standard is critical to the product. Thus the value of a standard is the value of that product line, no matter what the standard itself costs. The only reduction in value a standard undergoes by being freely available is to those who base their business on selling access to the standard.
To use a standard as an example, the SGML standard is seldom consulted by browser vendors - even large companies. As a result, so many browsers had bugs that would have been avoided if the standard had not been ignored until a new, open standard replaced it on the web. Thus, this standard has a higher cost and a lower value than the free standards used by the industry. This is exactly the opposite of what Mr. Bank predicts. It also means that Mr. Bank's message about the evils of free standards is being delivered with technology built almost entirely on free standards.
Finally, the most popular application of the SGML standard is HTML. HTML became popular largely because the standard is free, making it possible for individuals to create both browsers and documents with little or no cost.
Thus we see that free standards can be produced without funding from charging for access, and that charging fees for standards prevents their use and stifles innovation not in some subsidiary industry, but in the ones directly affected by the standards. Free standards, on the other hand, lower the cost of entry into a market, and encourage the global adoption of the standard - which increase their value by increasing the market.
Thus, to the people who use them, free standards have more value than commercial standards. Contrary to what Mr. Bank believes, Thomas Jefferson knew this to be true. He said not merely that information should be given away for free, but that there should be no legal protection for people wanting to restrict its free flow. He stated the case better than I could in a letter to Isaac MacPherson:
Free standards in particular bring more benefit to those who benefit most - the consumer of products that conform to those standards. Free standards may seriously damage companies whose business model is built around charging for standards, just as the automobile seriously damaged companies whose business model was built around the horse-drawn carriage. That is no reason to deny the benefits of free standards to the rest of us.