Minimising the muddle

Peggie Rimmer was Tim Berners-Lee’s hierarchical "leader" in CERN’s Data and Documents division when he invented the World Wide Web


Reams of material have been written about where, why and when the World Wide Web was born, but what about its conception? Gestation was rather like that of an elephant, difficult to know it had started and taking almost two years to complete. In fact I think the title of Tim Berners-Lee’s book, Weaving the Web, published in 1999 with Tim dubbed the inventor, is a better metaphor. When do a spider’s first few threads become a web? And when, if ever, is the job finished?

In 1984, Tim was recruited by CERN’s Data and Documents DD division and he elected to join the Read-Out Architecture RA section in the On-Line Computing OC group. I was the RA section leader and Tim worked with (and without!) me for the next 6 years. Mike Sendall, the OC group leader, agreed our work plans and held our purse strings.

At the time, CERN hosted lots of small and medium sized experiments using a variety of minicomputers, personal computers, operating systems, programming languages and network links. Back at the ranch, the OC group was endeavouring to provide data acquisition systems, the software used by equipment closely connected to the detectors, for as many experiments as possible. The conundrum, as in other areas, was how to embrace heterogeneity without having squads of workers generating exclusive solutions to intrinsically identical problems for bewildered users. Just the kind of anarchic jumble that Tim found challenging.

Several of us believed that standardisation, where apt, reduced waste and frustration. But the s-word was anathema in some corners of CERN, on the grounds that it stifled creativity, and we evangelists incurred the wrath of a few mandarins. Yet conformity seemed to rankle less when it came to electronics. Commercial companies were already competitively producing computer interfacing hardware that conformed to ANSI/IEEE international standards.

Hurrah! If you know the hardware you’re going to get you can prescribe how to handle it. I had worked with the NIM (USA) / ESONE (Europe) group that defined standard software routines for CAMAC interfacing and was on the committee developing hardware and software standards for the speedier FASTBUS system. Tim arrived as we were dotting the Is and crossing the Ts of the FASTBUS routines.

He was obviously a very smart young man (smart-clever rather than smart-sartorial!), full of fizz and, as a bonus, entirely likeable. When he presented his ideas in our section meetings, few of us if any could understand what he was talking about. His brain would overtake his voice, and holding up signs saying ‘Tim, slow down’ rarely had the desired effect.  We sometimes asked him to put things in writing, which didn’t necessarily help either. One of his erstwhile colleagues recalls ‘we knew it was probably exciting, maybe even important, but that it could take hours to figure out’. Listening to one of Tim’s presentations today one can still detect the run-away style, even after his intensive training in public speaking. However I do remember an occasion when his delivery was impeccable, in a play performed by the Geneva English Drama Society!

Tim’s main activity in the RA section was his Remote Procedure Call RPC, whereby a program on one computer could transparently access procedures, routines, on other computers even if they used different operating systems and programming languages, and whatever the network connecting them. He wasn’t too pleased when I asked him to specify the FORTRAN binding for FASTBUS routines, that is to define precisely the properties of the routines’ parameters as seen from within a FORTRAN program. Only later did he appreciate the value of that unwelcome task, when preparing the standards that would underpin the first two Ws of WWW. He knew that the job, however tedious, had to be done and done well, with the devil lurking in the nit-picking details.

Came 1990, another CERN reshuffle and Tim stayed behind in the new Computing and Networks CN division while the rest of us went off to Electronics and Computing for Physics ECP. Shortly afterwards I drifted away from ECP but I will always retain happy memories of the ‘80s and the pleasure of having Tim in our section. He was not the only singular character in that multifaceted team, but with his congenial personality he could work with anyone. At least I don’t recall having to field any complaints, apart from "what on earth is Tim proposing?" Well, now we know.