Microsoft & VRML

(This column first appeared in the April, 1996 issue of CAD++VRML)

This Column

If you’ve been reading this newsletter for some time, you may be familiar with my Garage Entrepreneur columns, through which I have tried to share my view of entrepreneurialism with you. If you enjoyed those columns and would like to see future ones, please drop me a note via e-mail.

In the meantime, expect to see this column, The Richter Scale, in this newsletter regularly. While the title of this column generally refers to the measurement of seismic activity, in this case it deals with my view of the world. In some Richter Scale columns I’ll provide opinions you may or may not agree with, and in others I’ll review products and offerings from various technology companies and rate them on a scale of 1.0 (pathetic) to 10.0 (earth shattering).

Microsoft & VRML
In this month’s installment of The Richter Scale, I’ll be taking a look at Microsoft’s role in the future of 3D model and world distribution, in particular with reference to VRML.

The Birth of ActiveVRML
As some of you may know, last December Microsoft released a specification for a new 3D description language called ActiveVRML – a name which many considered to be an act of hubris. After all, VRML is the name for a standard that has been developed by an industry wide group of 3D evangelists, and anyone, particularly Microsoft, using those four sacred initials in naming a product is equivalent to a virtual slap in the face. Especially since ActiveVRML doesn’t have much in the way of real VRML support, according to a number of vocal parties.

In questioning members of the www-vrml mailing list, no one seems to know if anyone actually owns the VRML name, something that’s not too difficult to believe of a group that deals more with collaborative development efforts than intellectual property protection. Unfortunately, that only makes it easier for corporate behemoths like Microsoft to trample their rights and efforts. My recommendation to any members of the VAG would be to immediately trademark “VRML” and send a letter to Microsoft to change the name of its product. Otherwise Microsoft may decide to lay claim to the name.

In any event, we’ve now learned that the “Active” part of the Microsoft specification name is part of an overall plan that Microsoft has to name similar technologies ActiveThis or ActiveThat, as in ActiveMovie and ActiveX Controls.

Microsoft Stumbles, Again
After calling their specification ActiveVRML, and submitting it to the VAG for consideration for becoming VRML 2.0, the next act of arrogance Microsoft committed was to not participate in the industry wide discussion pertaining to what will ultimately become VRML 2.0. That is, until a couple of weeks ago.

At that point, Salim AbiEzzi, product manager for ActiveVRML (and coincidentally, a college friend of mine from more than a decade ago), dropped several messages on the www-vrml mailing list. The first was to denounce the marketing efforts by SGI and Netscape in promoting Moving Worlds as the leading choice for VRML 2.0. In Salim’s words: “Unfortunately, the future of VRML became the subject of a public relations campaign and not of a technical exchange in pursuit of the better answer.”

Boo hoo. For once, Microsoft’s fabled marketing engine was out marketed. The issue of whether ActiveVRML is technically better than Moving Worlds has been debated vociferously, but without Microsoft coming out the clear winner or clear loser.

In addition to the above statement, Salim provided a technical rebuttal to SGI’s comparison of Moving Worlds with ActiveVRML.

Unfortunately, with the polling for VRML 2.0 support just a scant week or so later, it was a case of too little too late. Salim and other Microsoft staffers were openly chastised for not having participated in the open discussions that had been going on in the months since ActiveVRML was announced and the time Microsoft decided to counter SGI’s claims.

In Microsoft’s defense, the company is not used to having to justify itself to anyone. In the PC industry, which many of the VRML participants are not part of, when Microsoft issues a specification, folks rally to Microsoft’s banner, paying as much homage as they can. In the case of the ActiveVRML announcement, I gather they kept waiting for someone to show, and that just didn’t happen. By the time they apparently realized it, SGI and Netscape had stolen their thunder.

As part of the damage control, Microsoft let www-vrml participants to download an alpha version of the ActiveVRML browser, but that required downloading a beta 12MB ActiveX SDK, as well as many more megabytes of a beta DirectX driver set. Those few brave soles who had enough patience to download all this stuff (I got about 2/3rds of the way through after 4 days of trying) couldn’t get it to work. All this in the last few days before the VRML 2.0 polls closed.

As part of the discussion Salim initiated, it came out that Microsoft has no intention of dropping ActiveVRML in favor of anything the VAG might approve as VRML 2.0. Quite to the contrary, Microsoft intends to market ActiveVRML, assumedly against VRML 2.0.

Microsoft’s Motivations
Why would Microsoft want to go against industry consensus? Well, it all comes down to control. Control is a concept that Microsoft’s Chairman, Bill Gates, has imbued his company with. Microsoft has learned through past lessons, control is the key to the future. If Microsoft can’t control standards, then it can’t maintain an edge over its ever diminishing base of competitors.

At a conference I was at a couple of years ago, I asked a panel of interactive television proponents if they believed in standards. The spokesman from 3DO piped up and said “Sure, 3DO believes in market standards… providing we own them”.

Microsoft has followed the same philosophy, a philosophy, that while critical for self preservation, is one that also tends to alienate socialist technologists (most anyone that believes in committee and consensus driven standards).

Perhaps the alienation wouldn’t be as extreme if Microsoft didn’t bully people into submission. However, the alternative would be working in a committee, which would give someone else that isn’t Microsoft control, and that just wouldn’t work for them.

I have first-hand experience with Microsoft’s ability to derail committee-based standardization efforts when such efforts threaten the future as they perceive it to be. Back in the fall of 1993, I was vice-chairman of the Video Electronics Standards Assocation (VESA), and had also kicked off the VESA Advanced Graphics Interface (VAGI) committee. VAGI’s mission was to create a cross- platform, royalty free, low-level 3D API and implement it on a number of reference platforms. Shortly after VAGI got its start, Microsoft charged in, presented OpenGL as its solution to 3D and told us not to bother with our efforts, since OpenGL already did everything we needed. When that didn’t fly, Microsoft offered to take everyone’s input on a true low-level interface, and produce something that would make everyone happy. They would integrate the result into what would ultimate be called 3D-DDI.

A number of VAGI participants jumped ship at that point, and that was the start of the end of VAGI, which folded the following summer, after a specification was actually close to being completed. The death knell was that not enough companies would commit the necessary resources to implement the specification, as the appropriate resources were, in many cases, committed to support 3D-DDI.

The irony is that 3D-DDI has come and gone, to be replaced by Direct3D, which as of this writing still isn’t shipping. However, it took the purchase of a leading 3D API company, RenderMorphics, to get Microsoft reoriented towards what VAGI had been working towards in the first place. The only difference? Microsoft controls Direct3D.

When it comes to Microsoft, control is key. If they express interest in some technology, but don’t seek to control it, either directly or by introducing competing technology, then they aren’t truly interested. If they are truly interested, they will seek to control.

But, Microsoft isn’t always successful. Look at PenWindows, Bob, 3D-DDI, and many other mis-steps they’ve taken over the year. However, you can’t say Microsoft hasn’t learned from those experiences.

Will ActiveVRML be a Bob or a Windows 95? Time will tell.