This is the second installment in a three-part series. Part 1 outlines the discussions that surround the evolution of the Linux kernel. Part 2 takes a look at the current state of opinions on the standardization process.
When Microsoft won its bid to make Office Open XML an international standard last year, it was a pivotal moment for many in the FOSS community and beyond.
The process had been a highly contentious one, with protests from nations and corporations around the globe, and the International Organization for Standardization’s (ISO’s) final decision was met with considerable shock, disbelief and even outrage on the part of some.
Cynicism on the topic persists to this day, and debates can still be heard on the question of whether the standardization process is fundamentally broken. At the heart of it all lies one central question: Have standards become nothing more than a way to achieve vendor lock-in, or does openness still stand a fighting chance?
‘That All Looked Nasty’
“I stay away from standards bodies,” Linus Torvalds, creator of the Linux kernel, told LinuxInsider. “That seems to be a whole rat’s nest of backstabbing and politics, and I really don’t want to have anything to do with it.”
As for the OOXML case, “anybody involved with that must have been crazy,” he added. “That all just looked nasty.”
Such standards are not even the ones that end up being most important, Torvalds asserted.
‘Life Just Isn’t Long Enough’
“I’m personally of the fairly strong opinion that the standards that tend to really matter are the so-called de facto standards which may have been ratified, but were done so after the fact, rather than up front,” Torvalds explained. “The original POSIX.1 standard comes to mind — it turned out to be a pretty good and successful one, but largely because it mostly codified existing practice, rather than try to make up a new one.”
Regarding the intersection between openness and standards, “openness in itself ends up fostering standards, rather than necessarily the other way around,” he said.
Torvalds doesn’t spend too much time worrying about the question, however: “My job is to do the best dang technology I can — or rather, to smooth the way so that others can,” he explained. “I simply cannot be bothered to worry about standards bodies and vendor politics, etc. Life just isn’t long enough.”
True Standards vs. ‘Marauding’ Impostors
A true standard is “open, not subject to heavy-handed litigation, widely used — voluntarily, easy to use and adaptable,” Elbert Hannah, coauthor of the O’Reilly book, Learning the vi and Vim Editors, told LinuxInsider.
Current examples — at least in practice, if not strictly speaking — include MP3, TCP/IP, vi/vim for editing, and POSIX, he agreed.
On the other hand, some examples of “non-standards marauding as standards” include Word, Windows, DOS, Intel chips and generic HD DVD, Hannah asserted. “The above are examples of things everyone uses, but were not widely adopted on merit but instead by leverage of powerful marketing presence,” he explained. “They are not necessarily good or bad — they just never got to be voted on.”
‘This Does Not a Standard Make’
It’s not required that standards be free of owners, “but it is absolute that if someone owns a standard they don’t have absolute power and control — i.e., they can’t dictate how the universe uses the standard,” he stressed. “DARPA has been good with this; others haven’t.”
Today, standards “seem to be more about locking in vendors with customers, Microsoft’s OOXML debacle being a good example,” Hannah asserted. “Microsoft claimed interest in an open and unified standard for documents, but under the covers, their idea of a standard skews heavily towards their software and platform.
“This does not a standard make,” he said.
“I like the original model: Find a need, solve it, and let it be widely adopted,” Hannah concluded. “It’s important that for this to be ‘real,’ it has to be without undue influence from any interested third party. Unfortunately, today powerful vendors push suspect ‘standards’ to the detriment of open and friendly technology.”
Indeed, while “the wonderful thing about standards is that there are so many of them, the problem with standards is that they can be mandated,” Martin Espinoza, a blogger on Slashdot, told LinuxInsider.
It’s important to remember, however, that “Office documents were the de facto standard before there actually was one,” Espinoza pointed out. “What the ISO is doing is codifying corruption. By and large, the citizenry does not care about very large lies. Comparatively small lies like this one are uninteresting.”
In any case, Microsoft may have won the battle, but it lost the war, Montreal consultant and Slashdot blogger Gerhard Mack told LinuxInsider.
“I’m sure that they thought they would have a new standard that they were light-years ahead of the competition in supporting,” Mack explained. “The reality is that didn’t happen.”
‘Not Since the Inquisition’
Instead, “OOXML got changed during the standards process so now there are no known products on the market that support it” — including Microsoft’s own Office 2008, he noted.
This, of course, “is on top of all the bad will they generated by trying to stack committees in their favor and attacking dissenting voices with a fervor not seen since the inquisition,” he added.
For the moment, at least, “the standards process hinders openness,” Mack asserted. “Unfortunately, standards have become a tactical weapon that corporations use to their advantage. We saw the results of that thinking with the OOXML push by Microsoft and when Rambus co-opted a standards process so that the standards required use of their patents.”
‘They Might Find Themselves Irrelevant’
Looking ahead, standards organizations “need to sit down and admit there is a problem, and then decide if they want to help or hinder industry progress,” Mack said.
“I’m hoping the ISO recovers from this,” he added. “They have been publicly humiliated by the debacle but refuse to admit anything went wrong. Monty Python’s ‘Black Knight’ imitation would actually be amusing if the stakes weren’t so high.”
The group must not only “rethink their fast-track process,” but also “they need to create a standard for how member countries select representatives,” he concluded. Otherwise, “they may very well find themselves rendered irrelevant and pushed aside by people who need standards to actually foster interoperability.”
FOSS Debates, Part 1: Kernel Truths