Opinion Piece

This document is now old news. In August, a compromise was worked out whereby Microsoft and W3C backed down gracefully, we got a token Microsoft co-editor for the XML specification, and everybody claimed that's what they wanted all along. So this is an interesting story, but may be interpreted as a demonstration that the process works. Or not.

On June 27, 1997, I was dismissed from the co-editorship of the XML language and hyperlinking specifications by the management of the World Wide Web Consortium. Thereafter, Jon Bosak, the chair of the XML activity, acting on advice from W3C management that an objection from Microsoft to my continuing in that role was to be considered substantive, ruled that he could not properly designate me as co-editor. Microsoft's objection was based on the fact that I am, on a consulting basis, representing Netscape's interests in the XML standardization process.

Note Inserted July 23rd
The XML Working Group has been informed by Jon Bosak that either the W3C's position has changed, or he was mistaken in interpreting it; they are not currently decreeing that Microsoft's objection must be taken as substantive. The decision as to whether I should contine in the XML co-editorship remains outstanding at this point in time. This information is material, and changes at least the appearance of the W3C's role; accordingly, this piece has been significantly updated.

This note is an opinion piece outlining my view of the meaning of this sequence of events. The opinions are strictly my own, and not those of any other position or organization.

My Role

For the record, I was eager to continue in my role as co-editor, and have explicitly stated my willingness to commit to making enough time available for the forseeable future, whether or not the Netscape relationship continues. I have enjoyed this work immensely; enough, in fact, to have done it on an unpaid basis between September '96 and now. The agreement with Netscape, while better than not being paid, does not come close to paying for all the time that goes into XML.

Editing XML documents is difficult; it requires acting as the voice of a committee. Given that the committee rarely achieves perfect consensus, the editors are regularly forced to put things in the spec with which they violently disagree. The relationship with Netscape does not, in my view, change this; no amount of money from anyone could increase the difficulty here. The reason the editorship is bearable is that in a surprising number of cases, decisions that at the time looked bad have started to look good in retrospect; in other words, the process seems to work.

The suggestion that money from Netscape would damage this process is ridiculous, and to their credit, nobody who has actually read the spec or observed the process seems to be making it.

The W3C

W3C's key principle is that it is supposed to operate on a basis of consensus, defined as the absence of substantive objections. There could only be two bases for a decision to regard Microsoft's objection as substantive:

  1. a general policy that representatives from Netscape or Microsoft are not eligible to participate (aside from committee membership) in the work of creating W3C specifications, for fear of conflict of interest.
  2. a specific fear that in this case, Netscape's payments to me will cause me to slant either the contents of the spec or its public positioning in Netscape's favor.

Everyone is at pains to assure me that #2 is not the case, and I tend to believe them. Thus we are left with the conclusion that policy #1 is in effect. Since no such policy appears in any W3C process document, it can only have been arrived at by W3C management fiat.

Such a policy would be misguided at best. Here are some of the reasons:

  1. Ruling out contributions of personnel from the W3C's best-informed and most energetic members is a severe limitation on the ability of the consortium to produce quality standards in a timely fashion.
  2. It creates a two-tier hierarchy in the W3C: major browser vendors and everyone else. This is wrong.
  3. It creates an adversarial atmosphere; if employment by any Consortium member is assumed to be such a corrupting influence that employees and even consultants cannot be trusted to avoid abusing the process, what hope is there for building consensus on open standards?
  4. The policy is without precedent. SGML and SQL are examples of standards that were substantially authored by employees of companies with deep and direct interests in the technology.
  5. The policy is self-perpetuating; if certain companies are excluded from office on the basis that they might sin, there can never be a chance for them to prove their bona fides with unbiased service.

It should be noted that nobody denies the possibility that some member of the consortium might abuse an editorial or chairperson role to bend the process in their employer's favor, and that organizations such as the W3C must be vigilant in watching for such behavior and expeditious in stamping it out.

Microsoft

Microsoft has not said anything in a public forum about this maneuver. In private discussions it has been presented as a problem of perception; a worry that the press might get the wrong impression. This is ironic, given that between April 30th and mid-July Microsoft used their Web site to deliver the following lie, ignoring repeated requests from Jon Bosak to correct it:

Does XML have broad industry support?

Yes. Led by Microsoft, XML is supported by Sun, SoftQuad, NCSA, Hewlett-Packard, and now Netscape (as of April 1997), among other companies.

The most reasonable interpretation of Microsoft's demand for my head is that their interest in the standards process is as a vehicle to obtain tactical advantage, and that there is no demonstrable willingness on their part to work in a collegial and open fashion to make the Web a better place for its users.

This is a reasonable course of action for Microsoft, and it is hard to feel any sincere anger at them. Microsoft's core business model in recent years has been to achieve control of an ever-increasing proportion of the universe of software, until everything significant becomes "part of the Operating System" and part of the Microsoft revenue stream. If I were a Microsoft shareholder (and I have been) I would question how Microsoft can reconcile its highly successful business strategy with any substantive engagement in an open standards process.

Netscape

(This section was not written nor approved by anyone at Netscape, and represents only my impression based on the time I've spent talking to people there.)

In early 1997, Netscape was apparently disillusioned with the W3C standards process, and was not playing a very meaningful role in the W3C's efforts. We in the XML community felt this to be potentially damaging to XML's chances of acceptance, and Jon Bosak, Lauren Wood, Dave Hollander, and many others of us did an immense amount of work to break through this resistance. These efforts were eventually crowned with success, and (speaking, once again, unofficially) I detected a vastly increased willingness at Netscape to participate in a variety of W3C activities in an open and engaged way.

Following the events of the last week, the remarks I hear in the halls at Netscape are along the lines of "why bother with this process?" and "W3C is in Microsoft's pocket anyhow."

Whether or not these opinions are correct, I feel that Netscape's continued involvement in the process is good for both Netscape and the community. However, at the moment, I find myself rather short of good arguments.

XML and Related Standards

Despite the childish stupidity evidenced by recent events, I retain 100% of my faith in the power and utility of generalized markup, and in our work here to make it easy, lightweight, and accessible. Too often, the computing profession has discarded decades of process in favor of some proprietary hack job. This has happened, for example, in the domains of operating systems and user interfaces; we are trying to keep it from happening again in the area of document markup.

The People Involved

It should be said that the foot soldiers involved in this nasty little piece of trench warfare have done very well at keeping their heads and snatching some shreds of honor from this mess. Jon Bosak, unsupported by any written process, trapped between what feels like a contemptible power-play and what smells like fumbling cowardice, has managed a remarkable amount of transparency and steadiness. Jean Paoli, it seems, did his best to convince his management that this move had a downside. The rest of the WG has been kind and supportive.

As for the decision-makers at Microsoft who launched this shameful assault on the process, my opinions are obviously bent severely out of shape; anyone who cares should check the facts for themselves and draw their own conclusions.

- Tim Bray, July 1997