CERN Accelerating science

Abstract

A retrospective to events in the CERN Computing arena few years back.

Five years ago...

• End of DXCERN Ultrix Service (CNL issue 222 "Migration Special - Mail Issues" - January-March 1996)

After many years of faithful service the dxcern Ultrix service will be stopped towards the end of 1996... DIGITAL is encouraging users to migrate to Digital Unix...

This year was of particular relevance to CN services. It meant the bridging to new technologies and the stoppage of old hardware and operating systems. It was also two years away from the DEC-Compaq merger, which symbolically expressed the merger trend of the late nineties (Ed.)

• Roadmap to MAIL Services (CNL issue 222 "Migration Special - Mail Issues" - January-March 1996)

The main criteria you should use when choosing your preferred MAIL service are: the combination Desktop/OS with which you feel most comfortable; and the need to process your mail from more than one location...

This is not the only relevant article on mail, it was rather part of a dedicated CNL issue. I would tag 1996 as the year of our "User Desktop Revolution". Hundreds of users who had come to terms over the years with the console screen and rather unexciting interfaces, started encountering the richness of a "multiple-screen" environment. Not always as a more pleasant experience! (Ed.)

Ten years ago...

• World-Wide Web: On-line information for everyone (CNL issue 204 - October-December 1991)

A world of information is now available online from any computer platform. Information sources at CERN and across the world span subjects from poetry to biochemistry and supercomputing. We summarise the information currently sourced at CERN, and we introduce the WorldWideWeb (W3) program which allows you to browse and search all the data in a simple and consistent manner...

Ten years back a few of us were colleagues of Tim Berners-Lee. I was even in the next office and, for a while, his (nominal) supervisor. He had a very clear vision of the worth of http, and history has taken care of putting it at the right place.

This CNL article describes in some detail his line-mode interface, as such the proof of concept (enhanced by a Next-special variant). Few of us thought that an X/Motif interface would make a real difference in user experience for the UNIX-based physicists.

Unfortunately not enough effort couldd be invested in this aspect. Then NCSA/Mosaic came out; the rest is history... (Ed.)

• Can we and should we save the CNL? (CNL issue 205 - January-February 1992)

The objective of the CNL has always been to inform users about the computing infrastructure at CERN. The content and form... have been static. There was almost no feedback from the readers. We need feedback.

It was true ten years ago, it is true still today. No initiative meant to be useful to customers can survive without customer feedback. Luckily (for us) Bernd Pollermann did get enough feedback to continue, such that we can now celebrate its anniversary (Ed.)

Twenty years ago...

• 6250 BPI tapes on CDC (CNL issue 156 - January-February 1981)

Now that NOS/BE is running without problems ... 6250 bpi tapes can be used freely on the CDC machines. Since not all units are capable of processing tapes at 6250 (or 800) bpi, the parameter...
Use PARTIAL STAGING whenever possible.

6250 bpi tapes were a novelty in the early eighties, storing three times as much as a "normal" tape (about 150 MB worth of information). Part of the spectacle at any Computer Centre of this time was seeing the set of tape units all chewing data in jerky movements.

Nowadays that much can be stored in one "SuperDisk" diskette or a fraction of DVD, and it can be read at our home computer. We will have to find other excuses to be impressed (Ed.)

• Adding Black stripes to a car does not make it go faster (CNL issue 156 - January-February 1981)

A letter from one of our readers giving his opinion of FORTRAN 8X.

This is a criticism of the draft standard of FORTRAN 8X, summarised in the last CNL. Besides some technical questions, one particular point arises, namely, is it really worthwhile to "pump up" a well established language well beyond its limits?

FORTRAN 77 survived in the eighties and the nineties at least two attempts at "revamping". Meaning, of course, as a de-facto standard in High Energy Physics software. The last question remains: that is probably one good reason why Java was born rather than inventing C+++ (there are many others; however, I don't intend to launch a polemic now) (Ed.)

Thirty years ago and beyond...

• A Large European Computer installed at CERN (CNL issue 60 - 2nd April 1971)

The main computer for the Omega and Split Field Magnet (SFM) Projects, a CII 10070 manufactured in France by the Compagnie Internationale pour l'Informatique, was delivered to the Data Handling Division and installed in Lab.5 during the month of December 1970.

... The software includes a multiprogramming operating system produced by CII (SIRIS 7).

I guess that "multiprogramming" here refers to multitasking; I need a real computing historian here. Twenty years before anybody would offer a desktop alternative to UNIX.

But it is well known that one prefers to run onto one's own mistakes than to learn from the experience of the others; eh Bill? (Ed.)

• Programming Enquiry Office (CNL issue 60 - 2nd April 1971)

The Programming Enquiry Office is currently only staffed with two members due to transfer and leaving of staff. In May we expect one new member to arrive while two further posts are still vacant. Be patient, we will try to operate the office full time, but it will be difficult to offer the same service as before.

Wow! So not much of a change in thirty years!

When I read these statements one profession comes to my mind: public teachers, who are considered an essential foundation of this society but whose vocation is considered near to priesthood (and therefore social recognition goes accordingly).

We have a saying in Spanish which says "Itch taken with pleasure does not bother". Yes, it must be that :-)) (Ed.)