May 2003 Public Policy Computer Graphics Column

Introduction

Laurie Reinhart

First we describe our course proposals for the 2003 SIGGRAPH conference. Next, responding to changes in conference content and to a request from SIGGRAPH President Alain Chesnais, we made a proposal of how the Public Policy Program should fit into the structure of the annual conferences. The issues discussed in our tutorials and in our BOF's affect all computer users and need a designated place in the conference.

Broadband internet access becomes more and more important as both web use and as file sizes increase. Bob Ellis, with the assistance of Myles Losch, proposes a policy statement regarding broadband access.

Finally, our guest submission is from Sandy Ressler, member of the Board of Directors of the Web3D Consortium. He discusses intellectual property issues and the development of the Web3D Consortium Standards.

Policy Proposals for SIGGRAPH 2003 Conference

Bob Ellis

As in recent years we have submitted multiple proposals for policy events at the conference. We had to drop our plans for a panel proposal when it was announced that there would not be a panels program in 2003. I also decided not to submit a proposal for a third course (on broadband and digital television) primarily because of the work load required to develop the course if the proposal was accepted.

As a major expansion I decided it was time to expand the introductory course from a tutorial to a half-day course. Attendee response indicated that there was a strong desire to see more time devoted to the subject. We also submitted a proposal for a tutorial on copyright based on the 2002 tutorial. One of the speakers, Dan Burk, indicated an interest in adding software patents to the course. I felt the only way we could do that was to expand it to a half-day course. But the early feedback from the courses committee on adding software patents was not positive, so I dropped the idea.

We will also have our usual BOF and committee meeting, so once again we are planning on a significant presence at the conference.

Proposal for Policy Activities at the Annual Conference

Bob Ellis

In the SIGGRAPH 2002 summary in the November column, I mentioned that ACM SIGGRAPH President Alain Chesnais suggested I make a proposal regarding the place of policy activities at the annual conference. In response, the following proposal was submitted to the SIGGRAPH Executive and Conference Advisory Committees, but the EC decided not to impose a mandatory policy program on the conference.

SIGGRAPH Technology Program and the Annual Conference: A Proposal
Bob Ellis
November 2002

Summary

A proposal is presented to have the SIGGRAPH Technology Policy Program's activities at the annual conference be a joint responsibility of the EC and CAG. The EC would insure that there is room in the conference program for policy events and the CAG/Conference Committees would insure quality content of the policy events.

Background

The SIGGRAPH Technology Policy Program existed for several years without a significant presence at the annual conference. We held a BOF, a committee meeting and talked up policy, but we weren't represented in the formal program.

Recently this has changed. In 2001 we had a tutorial course and a panel. In 2002 we had two tutorial courses. All were highly successful in our terms. Attendance was small (100-150 attendees) by SIGGRAPH Conference standards, but the audience was enthusiastic and passionate about the issues. Another measure of success was that our annual BOFs that were held immediately after one of the formal program events saw attendance jump from earlier BOFs 1-2 to 25-30 people.

A Problem

But there's a problem. In 2002 proposals for a third course and a panel were rejected. The course rejection was due to space in the program and was not a content or quality issue. The panel proposal had a problem (no computer graphics panelists) and the rejection was reasonable. I've taken steps (bringing in a computer graphics veteran to serve as panel organizer) but the lack of a panel program for the 2003 conference made this move moot. (I'm looking at the 2003 Graphics Conversations venue but this probably isn't the sort of program that would be attractive to important panelists.)

So the problem is that the Technology Policy Program can generate more quality content that serves a small, but important group, in the CG community than there is room for in the conference. Also, the SIGGRAPH EC by supporting the Technology Policy Program has indicated that it feels this is an important topic. We can write columns for Computer Graphics forever, but nothing has the impact of having events on the formal conference program.

One aspect that isn't a problem is the requirement for proposals and review. This is essential to insure the quality.

Characteristics of Conference Policy Events

Technology Policy conference events should have several characteristics. These include: presentation as formal conference events, descriptions in appropriate conference materials (programs, publicity pieces, course notes, conference abstracts, etc.), distributed throughout the conference week and located in easily accessible rooms.

Policy events have modest space and AV requirements. Rooms capable of holding 100-200 people and the ability to show PowerPoint slides are about the only requirements.

The Proposal

As a long-time SIGGRAPH conference person (Co-chair SIGGRAPH '80, Vice-Chair for Conference Planning) I'm about the last person to suggest that the EC issue any but the broadest mandates for the conference program. But the Technology Policy Program is somewhat unique. It is an activity that attracts a small but enthusiastic and passionate audience. The SIGGRAPH EC believes it is an important function for the organization. It doesn't compete well for attention compared to events such as the special events of the 2002 conference (Yoda and Beyond: Creating the Digital Cast of Star Wars Episode II, The Fate of Play: Game Industry Revolutionaries Speak Out and Spider-Man: Behind the Mask)!

I propose that the EC (in consultation with the CAG) adopt a policy that mandates some level of technology policy be present in each year's conference program, subject to review of content for quality and appropriateness by the relevant conference committee. The policy should take into consideration the characteristics listed in the previous suggestion.

Such a policy would insure that the growing set of conference attendees interested in policy activities would find useful material in the conference program. It would also provide an important venue and publicity for SIGGRAPH's interest and support for policy activities. Review by conference committees would insure that the high standards of conference content would be maintained.

Draft Policy Statement on Broadband Internet Access

Bob Ellis

Because of the importance to computer graphics of widespread adoption of broadband Internet access and my frustration with policy makers who see the Internet primarily as a means for the delivery of entertainment services, I've been working on a policy statement. The goal of this activity is to provide USACM with direction and urge that it use its presence in Washington to bring this view to the attention of policy makers.

The availability and adoption of broadband (high-speed) Internet service is of critical importance to the computing and telecommunications industries and the country. Graphically intensive applications such as remote medical diagnosis, enhanced e-commerce, personal data exchanges (e.g., sharing family photos and videos), delivery of entertainment and gaming, and augmented user interfaces are highly dependent on the widespread deployment of broadband.

Therefore, we suggest the following principles related to these services:

The Internet is primarily a communication medium:

The Internet is an essential medium for the delivery of electronic mail, privately produced visual material, commercial exchange, education/training, academic discourse and health, research, and political information. For example, the opinion of the special three-judge federal court in Philadelphia that first rejected the (1996) Communications Decency Act called the Internet "... the most participatory form of mass speech yet developed." -- as lately proved yet again by the popularity of the self-publishing web logging (blogging) phenomenon.

The "killer app" for broadband is not the delivery of movies:

We believe the application that will induce most people to move to broadband is the personal sharing of pictures (and eventually videos). With digital photography becoming widely available, many people would like to share their photos with friends and family. The easiest way to do this is by email attachments (as opposed to setting up a website, etc.), but a large set of photos would soon result in long uploads and downloads via dial-up, even when transmitted at typical display-screen resolutions. Of course many people would probably like to send high-resolution pictures so the recipients could print them. If this was successful, we believe sharing home videos would be next.

While the sharing of personally created visual material may be the application that gets individuals to initially adopt broadband, there are many other uses that will be discovered after broadband is widely adopted. Foremost among these may be the transmission of high-quality medical images to under served areas (after broadband implementation issues to these areas are solved). Other uses likely to be discovered after adoption include online/interactive gaming, accessing entertainment such as motion pictures, and education/training.

Security, privacy and reliability:

The always-on, high-speed nature of broadband Internet services make such connections attractive to people looking to break into a computer. In order for the public to be able to use these services without jeopardizing the security and privacy of their computers and personal data, it is critical that providers make available easy to use and understand facilities (e.g., firewalls) to protect their customers.

In keeping with the Internet's primary use as a medium for critical communication in many forms, such services must maintain high availability. Unlike entertainment, the key uses of the Internet are not optional and users will not utilize these services unless they are reliably available.

An important challenge is the development and use of quantifiable measures that provide customers and policy makers with a solid basis for comparing provider performance.

Affordability:

It is becoming clear that most potential customers will not pay a significant premium for broadband. Before there is widespread adoption, prices must be in the range of $20.00 or so per month. Current dial up customers who have a second phone line for Internet access might be willing to pay more because they would be able to discontinue the second phone line.

The best way to provide affordable services is to have meaningful competition among several (more than two) providers. The challenge here is to reduce regulation and constraints on providers while guaranteeing an appropriate level of consumer protection.

IP/DRM:

We believe that the music and motion picture industries current efforts in forcing adoption of technical and legal measures to protect their current business models will never be successful. Broadband offers many opportunities to adopt new business models that will not alienate their customers and provide healthy revenue and profits. This was proposed in a Dec. 2, 2002 editorial in the Wall Street Journal.

Internet deregulation and common carriage:

We believe that a competition-friendly environment is the best way to increase the availability and adoption of broadband Internet services. But the interests of the public must be maintained. In particular the right of the public to engage in protected speech and have their voices heard is critical.

Dial-up telephone service is designated as a common carrier service. This means that the providers must transmit in an expeditious and nondiscriminatory manner all messages presented to them. In exchange, the service providers have immunity from criminal and civil prosecution and lawsuits regarding the messages they carry.

We believe the solution to providing protection for speech over all Internet services is to make them common carrier services. Note that this may impact the current terms of service used by many service providers that attempt to forbid a broad range of unwanted online behavior. These terms include, but are not limited to: bans on "flaming", harassment, IPR violations, open mail relays, reverse engineering, bandwidth "hogging", malicious conduct, etc. Often, the supposed harm is to non-customers of that ISP. We believe some items (such as bandwidth limitations) could be built into the contract, while others (such as control over content) could not.

Note that common carriage does not mean that all customers would have to be treated equally. Different levels of service (such as bandwidth availability) could be provided based on price. But for any given level of service all customers (including information providers) would have to be treated equally.

Recent concerns about national security raise special issues. The investigation, prosecution, and punishment of those engaged in illegal activity by law enforcement acting under court supervision should be preserved while protecting the rights of law abiding citizens. ISPs should never hinder the delivery of a message or communication even if they suspect it to be associated with a criminal or potential terrorist act, unless the delivery of the message would result in physical injury or death to a person or persons or create economic harm. Nothing should mandate that ISPs use technical means to read messages (i.e. sniffers or search applications) to look for key words or phrases except under strict court supervision.

Intellectual Property Issues and the Web3D Consortium Standards Development

By Sandy Ressler
National Institute of Standards and Technology
and member of Board of Directors Web3D Consortium

Let's face it, Intellectual Property (IP) is a royal pain but seems to be necessary. First, note that the opinions expressed here are solely the author's and do not in any way represent the author's employer or the Web3D Consortium. In addition, any mention of commercial products or companies is not meant as an implied endorsement by the author or his employer.

In particular let's examine the role of IP issues as they effect standards development and subsequently the deployment of products which support the standard. The purpose of this article is to highlight some pitfalls when considering the inclusion of IP in formal standards and to discuss these issues in the context of one organization's (Web3D Consortium) history as viewed by an inside participant.

Some Brief History

The Web3D Consortium (originally known as the VRML Consortium) is the organization responsible for shepherding the VRML (Virtual Reality Modeling Language) standard through the ISO standards-making process. Many people (the "VRML community") were responsible for the development of VRML. Significantly, Silicon Graphics (SGI) put forward Open Inventor as a file format to serve as the basis of VRML,. Open Inventor was selected after an RFP process, . For an excellent history see "The Development of the VRML 97 International Standard by Rikk Carey, George Carson and Richard F. Puk (http://www.gscassociates.com/pubs/VRML_P1C.html) SGI's donation of Open Inventor IP was a major accelerator to VRML's development. Conversely SGI's ownership of IP (patents) for user interface aspects of CosmoPlayer (SGI's VRML Browser) was a stumbling block to the release of CosmoPlayer source into any type of Open Source licensing after SGI sold off it's interest in it's web graphics division (CosmoSoftware...RIP).

Recently X3D, the next generation web graphics standard, was forwarded to ISO for balloting as an FCD (Final Committee Draft). This is almost the final stage before completion as an ISO standard. X3D uses XML to encode the VRML scene graph and more importantly brings the power of the whole family of XML standards to bear on web based 3D graphics. Significantly there are no IP encumbrances in X3D.

Policy

The Web3D Consortium has consistently held to a rigorous policy of requiring contributors of technology to declare any potential IP encumbrances up front. Recently the W3C has gone through a lengthy process of debate and review and moved farther towards openness with a "Royalty-Free Patent Policy". During over a year of contentious debate the W3C considered making a major shift in policy by adopting a RAND (Reasonable And Non Discriminatory) patent framework. This approach was rejected in favor of a simpler more open framework. It is instructive to read the minutes from meetings of these discussions; for an example see W3C Working Draft of the work in progress at: http://www.w3.org/TR/2002/WD-patent-policy-20021114/

If the RAND policy had been adopted it would have opened up the possibility that patented technology could be part of official W3C standards and that licensees would have to pay. The Web3D Consortium officially (see http://web3d.org/aboutus/ipr.html) has a RAND policy but has yet to include any patented technology into its standards. (Note that W3C standards are called recommendations and do not have the legal imprimatur of ISO standards, although they are widely used.)

What is a Standard Anyway?

Strict IP policies that hold technologies close to the source may at first sound quite fair. Businesses invest resources and deserve to recoup their costs. Standards, however, are not products. Let's take a small step back and look at exactly what it is that standards organizations are trying to accomplish and what a standard is.

What is the purpose of a standard and why do companies and individuals expend their resources to participate in standards development? Individuals, not affiliated with large companies often participate in standards development for three reasons: a desire to contribute to the technical community, to participate in a project larger than an individual can do alone, and for the "fame". These are quite similar to the motivations expressed by people who participate in Open Source projects.

Companies send people to participate for business reasons that include getting a step ahead of the competition and making sure that technology moves in a direction compatible with corporate directions. If a company can get its IP included as part of a formal standard, all the better (from the corporate point of view).

Standards, generally, are specifications of open technologies. They are meant to help build an infrastructure upon which industries can thrive. The incorporation of company specific IP is clearly an impediment to the open nature of an infrastructure. Of course, infrastructure does not have to be open, it can be company specific, resulting in a monopoly or hegemony. History clearly shows that the more open an infrastructure the more likely it will be adopted. The open nature of networking standards such as TCP/IP and others resulted in the creation of the Internet. And the blossoming of the world wide web followed from the open nature of the http protocol and URL specification.

Conversely, look at the electronic book industry; it barely exists and a balkanized set of proprietary standards has prevented a large amount of content from being put into place. An interesting counter example is the near universal ubiquity of Macromedia's Flash and Adobe's PDF. Both of these formats met a consumer need and there was no standard to fill the functionality vacuum. SVG (Scalable Vector Graphics) is the evolving W3C standard that is playing in the same domain as Flash, but in my opinion, isn't a potential threat to Flash. Flash is a fabulous authoring tool and SWF is the Flash file format. In the ideal world one might see Flash continuing with SVG taking over as the file format. The advantage to a company like Macromedia would be the likelihood of more 3rd party products that can manipulate SVG increasing the value of the Flash authoring tool and broadening its appeal.

Finally standards have longevity and are suitable for archival storage. Standards are not subject to the financial misfortunes of companies. No matter who disappears or goes bankrupt, HTML, XML and X3D will remain viable specifications. Archival storage is incredibly important.

Downside

An example of maintaining a hard line against the inclusion of patented technologies were the events that occurred approximately four years ago with the submission by IBM of a geometry compression technology for VRML. IBM responded to a Request for Proposals from the VRML Consortium. One of the missing elements of VRML and X3D is the lack of a good quality geometry compression technology. IBM submitted a good technology that met the consortium's technical requirements; however the technology was patented and ultimately the two organizations never came to any type of agreement on how to include the technology in the standard. IBM in no way shoulders any blame for this; expectations for both sides were not clearly expressed up front.

When companies invest in technologies and obtain patents, as they should to satisfy their shareholders, they quite naturally expect to recoup some of their investments. However inclusion of their patented technology into an international standard could bring the companies a longer term benefit than simply a licensing deal. Inclusion of patented technology with royalty free terms would allow companies to have their technology included as part of an infrastructure and receive the "blessing" of the standards making organization, benefiting the company and the public.

The MPEG-4 Issue

One of the most difficult areas the Web3D Consortium has had to deal with is in its relationship with another ISO committee, the group developing MPEG (Motion Picture Experts Group). MPEG-4 is the next generation of video compression and distribution standards that have come from the phenomenally successful MPEG committee. The MPEG group has been responsible for MP3, which has revolutionized the music industry, and MPEG-2 which has made DVDs the most rapidly diffused technology in the history of consumer electronics. Clearly MPEG has had an immensely positive impact on the economy. MPEG, unlike VRML, contains many patented technologies. There is in fact a company, MPEG LA (MPEG License Authority http://www.mpegla.com/), whose business is to license and organize the patents inside the MPEG-2 and MPEG-4 standards. Make no mistake; this is legally very complicated stuff. The MPEG-4 Visual Patent Portfolio (as of Jan 1, 2003) contains 95 patents (in multiple countries) for 47 technologies. In 2002 MPEG-LA formulated a licensing policy for MPEG-4 which includes broadcast fees which caused a great deal of controversy (see "Anger greets MPEG-4 licensing scheme" http://www.eetimes.com/sys/news/OEG20020131S0061 ). Key contributors to the MPEG-4 standard, notably Apple and Sun Microsystems, vehemently objected to the policy but now everything seems to be sorted out. However, the long delay in formulating a licensing policy (nearly 2 years), left a very bad taste in the mouths of developers (see "Licensing decision ends MPEG-4 tiff": http://news.com.com/2100-1023-944051.html ) . There is currently an effort in a joint ITU/ISO group to develop a royalty free baseline for video, as a response to these problems (see "Terms of Reference for Joint Video Team (JVT) Activities" http://www.itu.int/ITU-T/studygroups/com16/jvt/JVTToR.pdf. It remains to be seen if the IP issues have been resolved enough to make MPEG-4 a success or not.

Going back to the issue of the Web3D Consortium, one might rightly ask what does video have to do with 3D graphics? As it turns out MPEG-4 also contains the ability to represent a 3D scene graph and it can represent content that places video on 3D surfaces, which should enable novel content. The original MPEG-4 3D scene graph was contributed to MPEG-4 from the Web3D Consortium, and is essentially the key Nodes (a node is the primary data structure in VRML files) of VRML. One issue the Web3D Consortium is concerned about is that MPEG-4 not simply take the VRML nodes, change the fields a little and then have someone slap a patent on it. This has not happened yet; however it is a concern. In addition some members of the Web3D Consortium are afraid to even look at the MPEG-4 documents out of concern that they will be legally "tainted" (have proprietary knowledge) which could potentially subject them and the VRML/X3D standards to legal action. All things considered it's an ugly situation. There is however some hope. One good point of intersection and for potential interoperability is via one portion of MPEG-4 called XMT. XMT is an XML encoding of BIFS (BInary Format for Scenes). There remains the potential for the creation of translators, via XSLT, to convert between X3D and XMT. Of course the underlying representations must be semantically compatible for this to work and while there is much in common, there has also been some divergence. If the marketplace demands some interoperability then it does seem likely that useful tools can and will be built.

Moving Forward

The path ahead seems to be getting clearer, at least for my brain. The inclusion of patented technology may be necessary to get the best technology. Standards however are not simply about codifying the best Technology; they are about creating enabling infrastructures. The open nature of standards is even more valuable than producing the best technology. In addition it is the customer's requirements and desire to use the standard, or not, that is the final arbiter. Open, royalty-free standards are the best approach to enabling infrastructures. In the long run this approach results in technology that has the potential for being widely adopted. It is the adoption and wide use of a standard that makes money for business not the inclusion of encumbered technologies.