CernDOC, a fertile ground for the web’s inception.

It is widely acknowledged that the World Wide Web took its first steps towards success at CERN. However, lesser well known is that earlier in the 80s, CERN teams had already developed CERNDOC, a very advanced documentation system, and one of the first to implement the client-server model. This same idea was later used in the development of the Web.


The scheme used by Tim Berners-Lee to present the web. In yellow, the CERNDOC box.

The CERNDOC initiative was pitched in the early 1980s by the DD division (now the IT division) as a solution for sharing and storing documentation produced by the physicists and engineers at CERN. At this point in time, the central computing facilities were based on IBM mainframes. Although PCs, Macintoshes and laser printers did not yet exist, network development was already active. “BITNET provided the networking capacity necessary for the IBM VM/CMS platform and it was employed by the CERNDOC project. The VM/CMS operating system provided users with private virtual machines on which they could create or retrieve documents that were stored on a documentation server”, recollects Eric Van Herwijnen, CERNDOC project leader. “Drawing on the client-server system developed at the Rutherford-Appleton Lab (Sysdoc), CERNDOC provided users with the ability to manage documents on their individual virtual machines”.

CERNDOC was a very powerful system for retrieving information but there were factors that impeded upon its ease of use: “Documents could only be retrieved by sending a request to the server and it did not include a browser interface”, says Eric. “CERNDOC was supported by a language that did not make use of hyperlinks, rendering it impossible to link documents together. One had to select documents from the database in order to view them ”.

"When we started to implement the system, we decided to use the Standard Generalized Markup Language (SGML)" explains Eric. "At the time, it was the language we were instructing people to use on our mainframes. It was a meta language that IBM had standardized for text processing. It did not directly contain processing instructions and it had to be translated to represent the text processing commands that it contained. CERNDOC pioneered the idea of storing documents using an application independent interchange format allowing them to be displayed on different client platforms.”
As a result of the complexity of the SGML language, it was not used for the Web. Rather, Tim Berners-Lee used features from a number of SGML implementations to create a more simplified language, the HTML, which included an anchor for hypertext links and did not depend on a Document Type Definition (see box).

As the Web started to gain a good reputation within the physics community, CERNDOC and SGML were faced with a big problem – the dawn of the PC era. "All of a sudden, PCs became much more popular than IBM VM/CMS," describes Eric. "Slowly, we moved first to Suns, then to PCs and Macintoshes. The world went “What You See is What You Get” thanks to Microsoft Word and Adobe PDF became the de facto document interchange format. You could say CERNDOC and SGML died at the same time."

Although CERNDOC did not meet a very successful fate, it contained ideas that proved to be an important stepping-stone in the path to the development of the WEB.

Document Type Definition

A Document Type Definition or a DTD is a required component when using an SGML-type meta language. It is a set of markup declarations that define a document type and provides instruction for which elements will appear in a document and where. XML is a simplified version of SGML.

by Jordan Juras