The First International World-Wide Web Conference was held in Geneva, Switzerland, May 25-28, 1994. For more background information, see the WWW '94 web page at CERN.
Arrival, meeting with folks at CERN
The keynote address was given by David Chaum from DigiCash. It was an incredible eye-opener. I'll summarize his talk with an excerpt from an article he wrote:
The growing amounts of information that different organizations collect about a person can be linked because all of them use the same key--tin the U.S. the social security number--to identify the individual in question. This identifier-based approach perforce trades off security against individual liberties. The more information that organizations have (whether the intent is to protect them from fraud or simply to target marketing efforts), the less privacy and control people retain.
Over the past eight years, my colleages and I at CWI (the Dutch nationally funded Center for Methematics and Computer Science in Amsterdam) have developed a new approach, based on fundamental theoretical and practical advances in cryptography, that makes this trade-off unnecessary. Transactions employing these techniques avoid the possiblity of fraud while maintaining the privacy of those who use them.David Chaum
Achieving Electronic Privacy
Scientific American, August 1992, pp. 96-101
His remarks colored many discussions throughout the conference. In the closing plenary panel, it was largely agreed that the Web needs a sort of "Bill of Rights" to protect the rights of citizens in the developing digital age.
Mr. Harding had quite an impressive collection of graphs of statistical evidence to support the conclusion: The Web is big, and it's growing exponentially. Any data that were inconsistent with exponential growth could be attributed to phenomena such as server crashes, national holidays, etc.
With regard to Mosaic, he mentioned some collaboration between NCSA, Adobe, and other parties on an API for interactions between "applets" on the client side (i.e. desktop).
It was an interesting talk, and I don't have time to do it justice here, but I will include a list of issues from his slides:
Currently, links between nodes on the web are constructed by the author for the use of the reader, but they convey little information to the machine that would allow it to do automated indexing or searching, or to make inferences.
There has long been a notion of "typed links" in the web software architecture, but it has not been deployed in practice. Relatively modest applications of this concept include links that associate a stylesheet with a document, or links that aggregate several nodes in a collection for the purpose or printing the aggregate document.
The success of Mosaic has brought the capability to "surf" the web to a huge user community. But to this community, the web is largely read-only.
Immediately, we can see that tools for distributed maintenance of the web would make life easier for information providers. But more importantly, the deployment of technology like Mosaic creates class distinctions in the web community: the community as a whole has grown tremendously, but the information providers remain a privileged minority, and if we are not careful, the views of that minority may misrepresent the actual views of the community that they represent.
Contrast this with USENET, where every newsreader has a "post" feature, which can creates an article with the same readership as the original article. This fosters free, open, democratic debate.
The Web was conceived as such a democratic forum, and it is only due to lack of resources that collaborative tools have not been deployed.
Tim Berners-Lee sees the future of the web as one in which the objects in the web represent objects in the real world such that a house-object might have an "owner" link to a person-object, such that changing that link to point to a different person-object has the effect of transferring legal ownership of the house between those persons. Obviously there are a range of security issues. But commerce on the web is coming, and other applications of virtual reality will follow.
The funding is not completely in place, and so to make a big announcement about W3O would be to steal the thunder of the upcoming announcement by the funding agencies. So W3O has not been formally announced.
But there were enough questions about what organization would carry the torch for WWW that Tim got permission from the powers that be to make some information available.
W3O will be an international consortium with two principal sites: one at MIT to oversee U.S. operations, and one at CERN to oversee European operations. The intent is that it will operate like the X Consortium: companies will pay a fee to be members, and in return, they'll get early access to the technology. The technical team at W3O will act as editors of the specifications, and will develop reference implementations. Contribution, discussion, and collaboration will be invited and encouraged.
This meeting was arranged rather haphazardly, and there were some important players missing. We met again on Thursday, but to summarize...
I presented some slides that a couple OLIAS project managers and I put together.
I explained that I began working on the HTML specification when I was in the documentation tools group at Convex Computer Corp. I was out of the loop for a year, but now I am on the OLIAS project at HaL Software Systems, where once again, I am grappling with the issues of delivering high quality products that interoperate with WWW, which is a moving target.
Recently, I began soliciting support for vendor-supported specifications for WWW technologies. Since it appears that W3O will handle these issues in the long term, I have narrowed my focus to achieving interoperability among products supporting HTML in the upcoming months.
We discussed some of the outstanding issues.
The vendors showing support at this meeting include:
<email@example.com>was very open in representing his company's support for WWW: the HTML specifcation effort, WWWLibrary development, etc.
<firstname.lastname@example.org>from the technical publications department of SCO represented his company's interest in this effort.
<email@example.com>. In all the rush, I seem to have forgotten the details of our discussions.
See also: the call for participation, a summary by Bert Bos.
Bert Bos moderated this discussion about how SGML plays in WWW. The attendees were mostly bigtime SGML publishers that were interested in delivering their wares online. Intellectual property rights and mechanisms and HTML+ features were the hot topics, along with stylesheets, Adobe PDF, and legacy documents.
Tim Berners-Lee organized a sort of pizza-dinner and brainstorming session for Wednesday evening, but it turned out to be a casual chat out in the courtyard over a few beers.
Rather than hashing over technical ideas, mostly we got to know each other a little better. To (mis)use a new term from this conference, we did some "social engineering."
The object of the game was to get a list of needed future developments, and prioritize them. We managed to come up with a list, but it seemed pointless to order them. Tim Berners-Lee put the list online in in his notes.
This time, the vendors represented were:
We started out talking about the HTML spec document that I have been working on, but twice we wandered off and started talking about the WWWLibrary code.
Mr. Harding suggested setting up a task force to coordinate development of the code -- distributing diffs, avoiding duplicated work, etc. Eric Sink, myself, Ari Luotonen, Henrik Frystyk, and some folks from NCSA were nominated. I'm not sure what's next for that group.
There was some discussion about long-term maintenance issues. Tim explained the status of W3O, and where he thinks it's headed.
Yuri Rubinski suggested that the HTML spec could be published as an SGML Open technical report. This looks like a good idea. We form a technical committee, review the document, and publish it. The committee retains editorial and intellectual property control over the document.
The document will be available for review on the web through CERN's server. I gave a pointer to the copy on the HaL server, but that's only temporary. We'll probably spin an IETF draft or two in the mean-time too.
This was a long one, with lots of spicy discussion.
One interesting development is that right now, HTML is compatible with disabled-access publishing techniques; i.e. blind people can read HTML documents. We must be careful that we don't lose this feature by adding too many visual presentation features to HTML.
More discussion topics that I don't have time to expand on...
See also: CERN's collection of photos of the panel
Each of the panelists was asked to briefly discuss the most important future developments on the Web.
The same issues came up in several of the speaker's discussions: