National Language Architecture
or twelve striving years
Thirtieth anniversary of SHARE Europe
A synthetical review by Bernard Chombart
Bernard passed away on 2012-02-21
©SHARE Europe (SEAS) originally published fall 1992.
Have you ever thought that when an idea is ready and that it corresponds to the solution to a precise need, it almost always simultaneously appear, at the edge of collective consciousness, within many different brains? These are likely amongst the most creative ones and leave other types of skills to other individuals belonging to the mass of people.
Whatever the application area, this phenomenon can be observed, thanks to the powerful communication means that caracterize our era. It probably has always existed, but has not been really emphasized yet, because of the tendency of most humans to appropriate to oneself any invention which is actually a mere discovery, simply linked to a context good enough for its emergence.
In 1864, aluminium is still a laboratory curiosity (it is much more expensive than silver!), when the Frenchman Sainte-Claire-Deville decides to address the issue of its industrial purification: emergence of a need of which the mass of industrial chemists were likely to be aware, but did not want to see, in front of the inherently difficult problem. But the solution dramatically erupted in 1886, simultaneously in France and in the USA : each one on its own, Héroult, born in 1863 in Thury-Harcourt (Calvados, France) and Hall, born the same year in Thompson (Ohio, USA), fine tuned the exact same method, without knowing each other : the method is the simple reduction of aluminium oxyde (bauxite), by means of electrolytic process. A trivial idea, indeed, as it is always easy to claim, after. But bauxite is as abundant, almost everywhere in nature, as it is easy to mine. It is obvious to imagine that such a patent has a considerable scope, because it makes the cost of pure metal very cheap. The immediate consequence happened very soon, as soon indeed that the news is published on each side of the Atlantic: both "inventors" sue the other one in court. The trial was for that time very noisy, and held with a lot of care, taking into account the threat to immense interests. And because of the quality of this trial's conclusion, the good faith of both parties has been irrefutably proven, as a consequence, one can a fortiori emphasize that the "double" discovery was simultaneous.
What would have happened, had the means of communications of the past Century been equal to those of today? Who really discovered the AIDS virus? The controversy will last for a while, one can bet on this (Note of translator: The Americans have now admitted that the French did discover AIDS before them). But in fact this is not very important. I postulate that any invention belongs to the whole world, as the stars of the sky.
In computer science as well one can always fight (and indeed many do, using copyrights and supposedly dissuasive laws) for paternity or ownership of software. But this is much more difficult to prove than in the case of concrete inventions, because, here, design remains a pure abstraction, once the media used, which are only simple packaging, are excluded.
In fact, beyond plagiarism, which of course also exists, everything seems to happen as if a mysterious signal would give birth to an emerging phenomenon, like for germination during the spring. From gray matter emerges the idea, a bit everywhere. And the first bud to appear has no particular intrinsic value, whatsoever.
That's why, among all names that will be cited here, it is neither desirable nor even possible to know who has first triggered the action, one of the most unique and most original of our Association, which from now on tries to protect cultures in the face of information processing techniques. Beyond motivations indeed much less disinterested, which still remain very topical.
Until very recently indeed, nobody cared about problems linked to character sets, so long as data processing was only the business of computer specialists. Everywhere in the world, Anglophones dominate, and swim on the linguistic ocean, happy and in a way somewhat imperialist. In most cases, multinational firms must use their language. The others are so small a minority that their needs are spread far to thinly in their own macrocosm. In the non-English-speaking world, one survived so-so, most of the time unconsciously, like a sick person who endures her fate, and who gets used to it progressively, even if death must be the only end. Of course at the beginning only upper case letters were available; but since the invention of the typewriter, an Anglo-Saxon invention, most people still believe that one must not put accents on upper case letters (except in a few rare countries). When it comes to languages other than English and data processing - worse even simple respect - there was nothing. I even once gathered the following answer from the manager of a major German scientific university computing centre: "instead of wasting your time to teach French to your computers, you'd better teach English to all your personnel". But I would bet that this friend changed his mind since he is now in the business world...
Nobody was I saying ? No. Some crazy and/or illuminated fanatics (still qualified like this today in different circles), encouraged by a few isolated searchers to whom some homage is due here, had actively prospected, in the last ten or twelve years, at least; and much before one thought about it for the first microcomputer. If it had not been far of these "burned heads", the present situation would be much worse than the one we know, even if we rightly think that all remains to be done, as soon as we have the curiosity or the need to at least use a word processing program. And I would not talk about searching a non-trivial data base, nor about sorting a directory that has some international characteristics...
Even if consciousness of all these problems has been scattered, because so many bodies had taken them into consideration, and for a while, a powerful will to denounce them and to pragmatically struggle against them, was necessary.
Let's nevertheless affirm, and loudly, that such an honour must be granted, as soon as 1979, to SEAS1, later known as SHARE Europe.
And why precisely SHARE Europe ? Simply because its members constitute a diversified mosaic, inherited from multiple heterogenous atavisms of old Europe. So, SEAS gave itself the moral duty to bear this particular concern, being conscious that nobody else, on the planet, would care for a while. The other large "sister" associations are for the most part not multilingual. G.U.I.D.E.2 Europe, historically, has always been much more fond of concrete issues, with immediate or short term pay back (while SEAS beginnings have been rather marked by pure technics) and had cleverly bypassed the problem for a while, by dividing itself into linguistic "regions". This has not though precluded it to officially support our action, since 1985.
One can easily perceive here, by the way, how we, computer professionals, have a tendency to process our problems ahead of those of our own customers... Let's recognize, objectively, that we often impose them our ways, instead of carefully listening to what they want - up to the point where we even pretend that they don't know themselves. That is certainly one of the reasons that have delayed the process we're talking about, more than any technical or financial consideration.
The original feature of our own campaign has been to apply the first pressures directly and officially coming from users of computerized systems, in front (mainly through IBM) of vendors themselves, but also, later, more and more, to a number of standards bodies. And now our friends from "sister" associations, have brought us in the last years, a sensitive synergy: our message, from a very small working group, amplified because it is supported by the whole SHARE Europe, and IUGC3 as well, our message is now heard! And it leads, from now on, to consequences that can be felt worldwide.
It's not useless to strongly emphasize, because a wide range of professionals themselves are not really conscious of this fact, that these pressures have already been very fruitful. They have attracted the attention of highest level of IBM's management. Managers have been led to seriously begin the development, in that area, of mechanisms which we could not have been hoped without such pressures, except in a daydreamer's head. This is, by the way, one of the main reasons for the very existence of our users groups.
In that quest, begun by SEAS, for consistent standardisation of character sets, and for acknowledgement of languages other than English in their adequate processing by manufacturers (headed of course by IBM, that paves the way), one can distinguish four distinct periods. The first three cover more than one third of SHARE Europe's total age.
First of all, up to 1981, people progressively became conscious of problems that technological advance in hardware and software development was going to bring, as soon as we would get out of the narrow frame constituted by the bare Latin alphabet. A significant fact is that one of the first triggers will have been the action of expatriated British citizens (Martin, Gardner), much more irritated by these problems than natives, already trained to somewhat resignation. But among the latter, some protesting and clairvoyant futurologists were heard too. These alarm signals were generated by some friends who were real users (Péteul, Mailhol, Kagan), and not necessarily technicians, or even SEAS members.
The emergence of a Gordian knot in these multiple problems was gradually perceived as soon as 1979, by pressing remarks on meeting evaluation forms, and by an official requirement (# 088-0680), paradoxically coming from the OS (nowadays MVS) Project. Later on, many presentations were given, the oldest of which can be traced in the Proceedings of STM804, held in Durham; entitled "User's experience with communications", it is due to Michael MARTIN, and talks about the difficulties he had, then, in Denmark, to deal with the peculiarities of the Danish Alphabet (and also of the language) in a data processing environment, and also about the difficulties he had to consistently use in this regard IBM software and printers. He strongly blames, particularly, the impossible task to avoid intolerable overcosts that users, submerged in non-American-English environments, must abnormally endure . And with humour, he (orally) concluded with a play on the official theme, "Mixing and matching" by changing it into "Fixing and patching". That is why (for minor history) this theme has been renamed later "Mixing but not matching".
At the end of the same year in Antwerpen, events increased in importance at WPM80, where we got, as an answer to the above requirement - but only on technical grounds - two IBM speeches (Mr.LOIZIDS and Mr. SCHWINGE), completing those of Michael MARTIN (again), Birgitt MÖLLENHOF, and Bernard CHOMBART. The SEAS Executive Board then decided to raise the issue at the "Top Concerns"5 level, while a delegation led by president Kurt BAUKNECHT went to IBM Europe Management to present it officially, with, among others, an explosive speech by Frédéric ROUX; and to propose the creation of a "Task Force" composed of SEAS members from different countries, including an IBM specialist, having sufficient representative authority.
The second period began with the creation by president Jens LYNGE, of the National Character Task Force (NCTF). This working group met for the first time on the 4th of May, 1982 in Noordwijkerhout, near Amsterdam. Its works lasted for three and a half years, thus covering more than the entire presidence term of Hagen HULTZSCH. During its numerous interim meetings and also during the normal SEAS meetings, was gathered a synthesis of individual work voluntarily assigned to each of the NCTF members. They were initially: Knud NIELSEN (DK, chairman), Bernard CHOMBART (F), Michael TäSCHNER (D), and Hubert ICKES (IBM USA, Kingston).
But the action was somewhat delayed because, almost at its beginnings (after two meetings), three members out of the first four were replaced: Knud Nielsen by Peter GARDNER (DK, new chairman, UK native), Michael Täschner by Günther KRYSMANSKI (D), and Hubert Ickes by Jerry ANDERSEN (USA, Kingston). Jerry is an old expert in character coding, since its very beginnings. He brought us a very significant contribution, knowing eveything about the evolution of character sets, past and present, and about likely choices in the future. We owe him a lot.
Then around the end of the task force work, two new members became active in the project: Anders BERGLUND (CH) and Klaus DAUBE (CH), both from the "Office Systems" project. Thanks to Klaus dynamism, the "SEAS NCTF white paper on national character, language and keyboard problems" took a concrete and efficient form, from the numerous and diverse elements, previously extensively elaborated. The White Paper was finally submitted by Peter Gardner to the SEAS Executive Board at the AM85 Conference in Zürich, Burkhard MERTENS being president.
This booklet was voluntarily limited to a descriptive work, in fact turning into a very detailed requirement, but made on a global basis, which did not pretend to go farther than everything that concerned the Latin alphabet and its extensions (the associated different character sets and languages). Neither did it try to suggest effective solutions to IBM. The working group was indeed perfectly conscious that it was necessary to go further at a later stage, but this first approach, essential to a large part of Europe, was appearing as a high priority emergency.
After the publishing of this White Paper, particulary inside IBM, but also outside of it, a formal answer was made by IBM's Keith AUSTIN, during SM86 in Heidelberg. It was recognizing the excellent quality of that work, qualified by the General Management of IBM as "remarkable", and the importance of this contribution from our Association. It was announcing essentially the intention of IBM to officially take this work into account, for all its future hardware as well as software development.
The third period paradoxically began with the disbanding of the NCTF, after the publishing of our first booklet. But a small group of committed persons felt the necessity to maintain a pressure on IBM, particularly within the "Office Systems" Project. Because, beyond sincere statements of intention, the essential issue for IBM, when a project is presented, always remains this one: « What about your business case? Where are your short-term customers?»
Coincidentally we have then been invited by the French "Ministè re de la Culture" to participate to an international symposium. It was organized in Tours, on behalf of the Council of Europe. This event, gathering some 42 nationalities, in March 1986, had as a theme "The language industries" and was proposed to bring water to the mill of "stakes on Europe". Our White Paper had been pointed out by organizers, who wished to use it as a base for concrete work, since the debates had nothing more than theoretical issues to propose, except a classical products presentation, due to an IBM Europe marketing contribution.
It would be exaggerated to claim that this meeting produced by itself more than political dialectics, mixed with mutual congratulations, or the usual underground actions targeting machiavelic goals. By the way, we never received its proceedings (probably because it was held just before the Department had a new minister appointed), if they were ever officially produced.
But during the colloquium, by chance, a delegate from the Québec Government, present in Paris for a totally different issue, Mr. Pierre CHOUINARD, had sollicited an invitation to satisfy his intellectual curiosity: and enthusiastically, near Ronsard's grave, he came to me to congratulate us for the quality of our works, and to ask if Québec could join our pressure group, because it was exactly the type of forum his government was looking for...
A few weeks later, I took the opportunity to hear that an IBM representative of the Toronto NLS6 laboratory was traveling across Europe to make a survey on linguistic complaints of our diverse countries, and that I could meet him, if I wished. I decided to take advantage of this opportunity, and to arrange to get to this rendez-vous together with Mr. Chouinard. Thus we met Denis GARNEAU, in a tower of "La Défense" in Paris, in Maurice GIROD's office, who was in charge of PAO7 products within IBM France management. We had no difficulty to convince Denis to do all what could be done for him to become a privileged correspondent of our working group (then dormant), while on its side, Mr. Chouinard would do what was necessary so that Québec would become a member of SEAS.
Following this splendid combination of circumstances, SEAS welcomed in Jersey its first user member from North America, Alain LA BONTÉ, and under his strong impulse - because he had come for no other reason - AM86 has seen the working group "wake up with a start", and so much better that IBM had officially delegated Denis Garneau, who became for all of us an excellent friend later, not only extremely competent, but furthermore very effective. We owe him a lot, too. The active presence of the "Toronto Lab" was a proof that Keith Austin's statement of intention was not mere wind.
Since this privileged moment, the working group is more and more active, during all this third period. In Montpellier (SM87), it asks to become a full Project. But the kind of issues discussed, sometimes confidential, which required close symbiotic work with IBM, can not very well take place in a Project, which is usually a relatively open forum. Finally, the Executive Board decides to create a "Deputy Project Manager" post, chaired by Alain La Bonté, around Anders Berglund who leads the "Office Systems" project, and Klaus Daube. Well, one still had a tendency to assimilate our concerns - always the same, targeted to all computer related issues, but with a cultural dimension - with mere office automation.
A little later, under Sverre JARP chairmanship, it was decided to relaunch this definitely undestroyable group, under a new official SEAS structure, an STWG8 led by Klaus Daube, who formed a marvelous team with Alain La Bonté, Denis Garneau and Jean-Pierre CABANIÉ, without forgeting Kurt NEUENSCHWANDER, the new manager of the "Office Systems" Project, as well as a few others. As for the editor of this story, who took part in it since its very beginning, he limits his presence, for now on, as a discrete éminence grise.
Encouraged by a technical survey within membership (with a participation rate and results beyond all expectations), the work continues and can be seen in numerous interventions during conferences, but also intensively in the STWG activities. Among others, the group presents to Robert BETTENDORF, then Director of NLS, during SM88, the famous "Davos requirement"9. The underlying rationale of this requirement was: SAA did not address at all the issue of internationalisation or National Language Support.
This action, after a productive dialogue in Toronto - thanks also to the help of SHARE Inc.10 and especially Edwin HART - will serve as a base to IBM for so-called CDRA11 structures; while a second document is in preparation among us, with a much more forceful shape, pushed by our friends and IBM partners themselves. At their head, let's hail Ted SASALA, who always supported the STWG since he became NLS director, who assured the cooperation of leading observers and specialists, and comes to visit us in a participative fashion.
Simultaneously SHARE Inc. comes in 1989 to join SEAS in its efforts, with the precious initiative of Ed Hart, the leader of the ÆCS12 Project, which (roughly) gave itself a mandate to steer within SHARE the sensitive issue of character sets, before becoming ÆILS, extending its activities to problems involving languages. In parallel, SEAS, which has become SHARE Europe in 1990, comes nearer to ISO13, (in which body Alain la Bonté is an official representative of Canada, and Ed Hart of the United States).
The STWG puts a rather strong voice into the quarrel that opposes the de facto standard UNICODE (developed and adopted jointly by many manufacturers) to the standard project that ISO is currently discussing under code number DIS14 10646, and which, no doubt, will be balloted and promoted very quickly. Right now, it seems that we have already got a promising consensus, ISO being likely to integrate UNICODE within standard project 10646. For non initiated people, this point is essential, because this universal standard will very certainly be the governing rule, in the future, of all aspects of worldwide information processing technology, as long as character sets are involved (until the next encounter with extra terrestrials, has written one member in a recent report).
At last, in September 1990, under the presidence of both SHARE Europe and the IUGC by Iain STINSON, is published near Paris, the birthplace of SEAS, in la Défense, the second SHARE Europe's White Paper, entitled NATIONAL LANGUAGE ARCHITECTURE, because it is not less than the advent of a global challenge to the information processing technology; as long as it is declared in the foreword: "Technology has the duty to promote culture, not to destroy it". Again, this document is mainly due to the remarkable working power of Klaus Daube, but it is also co-signed by Alain La Bonté, Jean-Pierre Cabanié, Bernard Chombart, Peter Gardner, Denis Garneau, Edwin Hart, Kurt Neuenschwander, Daniel Pellet, Otto Stolz and Johan van Wingen. This time, the text does not merely describe in a more detailed fashion, some of the problems raised in 1985, but goes beyond in proposing concrete solutions, most of which have been proved in experimental environments by one contributor or the other. Thus, the proposals of this new White Paper are certified as viable, as much as necessary, and consistent with the ongoing efforts in international standards issues.
What about the fourth period? Well it has just begun. Denis Garneau has released his "powers" to Marty MARCHYSHYN. Each working day brings its share of news, of hopes, and also of deceptions. As has already been said, almost everything remains to be built in this domain, where at first glance all seems to be impossible to realize. That's why it is essential that those of IBM who are responsible to pave the future be sure to count on our support in partnership, and also that they endure a stronger and stronger pressure from their customers, whom we are delegates. That's why also the small NLA STWG must be able to count on the warm support of all SHARE Europe members, who must consider our action, even if they have no time to spend on direct participation, as their personal business. With regard to this, our survey mentioned above has been encouraging, but more has to be done. A lot of aggressive resistence still has to be fought, particularly in investment issues raised by all manufacturers, IBM itself being in front.
Only to such an extent will we be able to irradiate, grow, and implicate much more member associations of the IUGC. Our president for the thirtiest anniversary, Barry LANDY, is by chance, polyglot and perfectly conscious of the strategic issue, for the short term as well as for the much longer term. Let's take advantage of this fact, and let's give him a strong power of expression among the other presidents, his colleagues. We must make sure that our message, which seems to have been well launched, be permanently completed, and extended as soon as possible in concrete realizations.
Happy anniversary, my friends !
Many thanks to Alain La Bonté and Denis Garneau for their help in translation.
4 STM: Spring Technical Meeting - WPM: Winter Project Meeting. Since 1982, SEAS holds only two meetings a year, where the program structure was generalized: The SM: Spring Meeting, and AM: Anniversary meeting, which falls in Autumn.
8 Special Technical Working Group - a structure made official in 1987, whose goal is to process important issues, in a well defined frame, and generally covering fields of several different SEAS Projects.
9 This requirement, with a generic scope independant of all products, is the first one to have been integrally accepted as it is. This constitutes a première in the dialogue between IBM an its customers.