by Carl Cargill, UniNews online Standards Columnist
The recent decision by Sun Microsystems to apply to the ISO/IEC Joint Technical Committee 1 (ISO/IEC JTC1) for recognition as a submitter of Publicly Available Specifications (PAS) is an interesting action that bears further examination. While the action has been both vilified and praised by the press, the most common reaction has been "Huh? What did they do? What does it mean? Why did they do it?" This column will examine the first and second questions - and will hazard some guesses as to why they did what they did.
The Publicly Available Specification concept was hatched several years ago in JTC1, a formally recognized Standards Developing Organization (SDO). (An SDO is recognized by ISO as a legitimate place for standardization to occur.) It was a belated recognition by SDOs that their formal standards creation and publication businesses were losing to competition from consortia and to consortia specifications. To counter this threat to both influence and revenue stream, the SDOs created a way that these organizations and their specifications could be brought into the formal fold. The PAS process (the solution to the problem) is a two step process designed for a consortium to 1) seek recognition by the SDOs that the consortium has an open process; and 2) recognition by the JTC1 of the goodness and openness of the consortium's specification.
There were several things that helped convince the SDOs create this process. Membership in SDOs was dropping. While no longer hemorrhaging members as they did several years ago, the SDOs were not gaining new and growing companies into their membership. Software companies, database companies, application providers, and Internet/Web companies were not (and are still not) joining the formal SDOs. During the time that membership and participation has been declining in these groups, the Internet Engineering Task Force has been growing at a phenomenal
(and possibly frightening) rate; the San Jose IETF meeting had nearly 2000 participants. (The IETF is a consortium, not an SDO. It has been accepted as a liaison organization by ISO, but is not seen as a peer organization by ISO.) The Object Management Group (OMG) is approaching 800 members and growing at a rate of two new members per week. Other consortia (such as the Open GIS Consortium) have appeared and are gaining membership and an ever increasing following. (To balance this picture somewhat, some consortia such as OSF and X/Open have merged, and others, such as UNIX International, have even died.)
While the loss of membership was bad enough, the SDOs were not getting the new techologies. Object technology looked poised to finally take off - and it went to OMG. The Web is derived from work done in the IETF, and spawned a new organization (the W3C) to help it get started. The interesting and fun things were happening in the consortia - and consortia were getting the lion's share of recognition and, usually, dues money. Additionally, the consortia weren't as open as SDOs and therefore didn't have as complex a process as did the SDOs, resulting in specifications that could be delivered more quickly and with fewer options to confuse the providers and users.It was to correct this situation - the loss of influence and the loss of revenue - that the PAS process was created. The SDOs reasoned that a consortium would consider participating in this process because they knew that all consortia really suffer from SDO envy, and really want to be an SDO.
It was to correct this situation - the loss of influence and the loss of revenue - that the PAS process was created. The SDOs reasoned that a consortium would consider participating in this process because they knew that all consortia really suffer from SDO envy, and really want to be an SDO. The harsh truth is that the last thing that the funding members of a consortium want is for the consortium to be like an SDO; if they'd wanted an SDO, they would have increased funding to the SDOs in which they were already participating. The SDOs were engaged in the Tinker Bell theory of standardization. They believed that if they wished hard enough, the situation would get better.
Well, it didn't. Over the past two years, only a handful of consortia have applied for PAS submitter status, and less than five specifications have actually gone through the process. Basically, the consortia didn't want to be registered as PAS providers, and, even more disturbing, their customers didn't seem to mind using specifications that weren't formal standards.
This, then, was the situation when Sun decided to apply for acceptance as a PAS submitter. I do not believe that it ever occurred to many of the members of JTC1 that a company might apply as a PAS submitter. By definition, a company does not develop its product specifications in a forum as open as that of a consortium, let alone an SDO. They are not compelled to consider negative comments by users; they are not required to agree to license their technology in an open and non-discriminatory fashion. They are also usually not required to allow public comment on revisions - or any of the series of things that consortia and SDOs do. Basically, a company's product development process is focused on giving it an unfair market advantage so that people will buy its products and not a competitors.
So, how can Sun justify its application? Basically, I believe that Sun really believes that everyone loves Java and that everyone believes that it is the "Answer". Additionally, because Sun believes in standardizing interfaces rather than standardizing implementations, this seems to fit into their model of what "open systems standardization" is all about. The interface standardization emphasis can be seen from the work that Sun did on the Public Windows Interface (PWI) in ECMA, where they created a standard around the Windows interface, rather than the Windows implementation. Unlike many of its competitors in the OS/Web arena, Sun is willing to open up its interfaces and compete on the skill of its engineers, rather than business tie-ins. Interface standardization means that Sun is willing to bet that its implementation of Java will be better than anyone else's - because if it is not, Sun will have lost its unique market advantage.
However, there are other consequences to Sun's actions. They have opened a way - possibly inadvertently - for other companies to also request PAS submitter status and submit their technologies to ISO for consideration. They also may have written an inelegant closing chapter for many consortia and SDOs, because if a company can use a PAS process to achieve standardization, there is no reason for it to waste time and money on SDOs and consortia. Sun seems to have taken the old maxim of "If you want something done right, do it yourself" to new extremes. It remains to be seen whether this is a brave new world or an overbold attempt. And this is something that the market - and only the market - can and will ultimately decide. But, in the meantime, it should be fun to watch.