In the coming months, network security and commercialisation issues, such as how to create paying services, will be the highest priority subjects for Internet development, says Olivier Muron, director of promotions and technology transfer for France’s INRIA National Institute of Research in Informatics and Automation. Up until a few weeks ago, INRIA was just another Internet service location and M Muron’s sentiment would have been just another Web user’s speculation. But following an agreement with CERN, the Centre for European Research into Nucleonics, in Geneva, INRIA, along with the Massachusetts Institute of Technology, is now responsible for supporting and managing the evolution of World Wide Web technology. CERN has been the centre of Web technology development since Tim Berners-Lee, a CERN researcher, invented the thing in 1989.
Elite
That the invention came from the elite international community of high-energy physics was not surprising, since it is a very geographically distributed and computer-literate community that has been using computer networks to distribute documents and for electronic mail for a long time, says Mike Sendall, secretary for CERN’s Large Hadron Collider experiments committee. The World Wide Web was a welcome development because it alleviated the physicists’ headache of using multiple incompatible networks around the world to exchange the voluminous test data that is critical to better understanding the behaviour of the universe. Having invented Web technology, CERN naturally developed as the center of the standardisation effort for the Net (the Internet). In the early days of the Net, it was entirely appropriate for us to be the focus of activity, Sendall says. Today, however, is an entirely different story. No longer is the Internet the privileged domain of a largely academic computer user audience. The Internet Society reported that new Net subscribers grew 22% in the fourth quarter of 1994 alone, while the .net domain, used largely by commercial service providers, increased by an astonishing 66%. At this rate, the society predicts the Net will count 100m hosts by 1998. The centre of gravity has shifted, if you like; the Net is now of much broader interest than just to our community, Sendall said, by way of explaining why CERN chose to disengage itself from Web development and promotion. It was a question of the amount of work there is to do and to what extent it maps onto CERN’s mission, said Sendall. He added that the approval in December of the collider project made it increasingly difficult for CERN to continue its historic Web role. At the time, Sendall was already in discussions with the Massachusetts Institute and the European Commission on establishing joint European-US support for the Web standards.
By Marsha Johnston
Both the EC and CERN wanted a strong European presence with the standard and INRIA seemed like a good bunch of people to take it over, he said. Indeed, INRIA has been the starting point for some of France’s hottest young information technology companies, including object database management system developer O2 Technology SA, C++ tools developer ILOG SA and Standardised Generalised Mark-up Language expert Grif SA. Aside from running its own professional Web service, INRIA developed 100 Years of Light in the Paintings of French Museums as a Web service for the Ministry of Culture and was retained by the European Community as a national host for advanced Web services in its Advanced Communications Technologies & Services. INRIA has done quite a lot of Web development. They have not played a big role in central standardisation issues, but they are quite experienced in networking and security issues, Sendall said, adding that those are the two most important areas for The Net at this stage. Jean-Francois Abramatic, INRIA’s director of development, says he and Massachusetts expect to conclude a memorandum of understanding on the details of their collaboration by the end of the month. There are no problems, everyth
ing is moving along as it should; everyone is full of good intentions that the uniform standards will be maintained, he said. No decisions have yet been made on how to split development duties, he said. Service support will be divided up more or less geographically to minimise our telecommunications charges, but it will be transparent to the user, he said. Broadly stated, however, INRIA and the Massachusetts Institute’s work will comprise four main areas: evolution of Web components, including the HTML HyperText Mark-up L anguage, the http and url client-server dial-up protocol, the universal residence locator; development of a reference code; supply of general information about the Web; and promotion of the technology and its distribution. Muron noted that ergonomics and further improvements to the HyperText Language, such as better means for handling maths formulae and text flows around illustrations, are also on the table for development. Sendall said CERN will continue Web developments that are targeted specifically at its own user community, rather than generally focused issues. We want to stay in the game for our own community because we see the Web as an essential tool in maintaining coherence in our large physics collaboration, he declared.
Computing power
Muron places great importance on the informational aspect of INRIA’s charter. We have to provide enough detailed information about the Web so that someone who doesn’t know anything about it could learn on their own how to use Web technology, he said. In the meantime, INRIA is planning for an increase in its current computing power. We already have a Web service, but it doesn’t get the same volume of access. To pass to the level of directory service consultation as CERN’s, which is automatically accessed by thousands of users every day, we need bigger workstations and increased network capacity, he said. Sendall is not worried that CERN’s departure will cause Web standards to evolve away from the requirements of its users. We use lots of other standards that we don’t define, he quipped.
