My Virtual Computer History Museum
Large part of my research career aimed on data distributed in multiple data bases, relational ones especially. I interacted with many distinguished researchers. Ambitious students helped with Master or Ph. D. Thesis. A lot of these activities took place in Europe, while I worked at INRIA. These pioneers helped to shape the everyday life today. Their achievements are a part of Computer History. They became largely forgotten today. Especially, those that ended up as gray literature. Hence are not or no more indexed by search engines. The goal of virtual exponents that follow, also called souvenirs, is to make the memory of all this seminal effort persistent.
In 1977, I joined SIRIUS pilot project at INRIA headed by Jean Le Bihan, since the creation of that project in 1976. Jean was great guy and modest like a few. Hard to believe, passed away at 70 only. Peace to his memory. The project goal was distributed databases. Budget was many millions francs. French state and private industry really tried a push into this new domain. As early as in 1978, we started an informal mainly European research group and mailing list of folks interested in the field. Termed Interest Group on Distributed Data (IGDD). The group soon had hundred plus participants. Among outside SIRIUS “founding fathers”, I remember especially Prof. Prof. Holler (Karlsruhe), Herbert Weber (Berlin), Van de Riet (Free U. of Amsterdam) and Fabio Schreiber (U. of Parma). We had great time. Winds of life pushed us in different directions. Wish one day they bring us together again.
Soon after, IGDD started international conferences and seminars. Here are the cover pages of a series of Seminars entitled Distributed Data Sharing Systems. The name was coined by my dear colleague & SIRUS friend Gerard Lelann. We organized the 1st event together. Gérard retired. Wonder what his whereabouts are.
2nd Seminar, in 1981, migrated to Amsterdam, invited by Van de Riet. For the 3rd one Fabio Schreiber was the Chair. Another place, another style. The classical Italian castle, he brought us to in Parma was a jewel. Altogether, the three events have discussed hundred plus contributions. These were seminars, i.e., workshops. We provisioned time for ample discussion. As souvenir, here are a few photos I managed to recover by chance, with folks who made all this happen (pdf). Prof. Van de Riet chairs the discussion, sitting right under the screen. Next to Prof. Serge Miranda (Toulouse). The ones I spoke about are listening in 1st rows. Unless Prof. Holler is instead at the extremity of the presidential table. Next to Prof. G. Schlageter (Karlsruhe Institute of Technology),Where am I?
1st Seminar on DDSS was preparatory to an international conference on the topic. As the usual difference between a conference and a seminar, the former was supposed to have stronger selection criteria etc.. Here is the cover page of that one, in 1980 in Paris. We called it 1st Intl. Symp. on DDBs & it was the 1st conference on this topic ever. This, ipso facto, makes it especially important event for our museum. You see above the cover page of the proceedings. The book itself is out-of-print for a while. Here is at least the conf. flyer (pdf), recalling these involved in the event. What the flyer does not say, is that the person behind the conference secretariat was one of the best professionals I met, Miss Therese Bricheteau from IRIA. She helped in other events, including our DDSS seminar. The Program Chair was Prof. Delobel, Orsay U. Also great guy & friend by that time. See us little later in 85 at IFIP conf. in Japan (pdf). Time passed. He now seems gentleman farmer in France.
The symposium had demo session. The demos have shown a number of prototypes in development by SIRIUS project. Here are some recovered amateur pictures (pdf). If you zoom, you’ll see the names. From my chapel, you’ll see first ever multidatabase prototypes. These were heterogeneous information retrieval systems, termed Sysidore and Messidor. The latter became better known. Although, retrospectively the former was perhaps more realistic, given the current practice. Messidor was actually built on the original PC, predating the IBM one. Termed Micral (Bull). Better design than the Big Brother, but still did not make it & finished sadly (see the actual computer history somewhere). For Messidor, see publications through my home page. See also the souvenir photos from the demo by the creator of the system, Eng. Catherine Moulinoux, among exponents we present below. SMGB means MDBS in French.
The event was so successful that a 2nd Symposium was organized two years later by Prof. Hans J. Schneider. In Berlin, I believe from memory. I lost the proceedings somewhere, unfortunately. Keep searching. Feel free to send a picture if you have the book still. Fortunately, there is a copy of the table of contents at least on the Web pdf.
Since 1978, we continued in SIRUS the direction we have proposed by then and called multidatabase systems. By accent on autonomy, the concept was somehow close to that of federated databases, proposed by Prof. Mc Leod (USC), in about the same time. There were nevertheless important basic differences. Anyway, to recall, an MDB system manages autonomous but also, as we said, interoperable databases. The user perceives the existence of several databases, perhaps with various semantic differences. The system offers the capabilities for convenient queries addressing jointly data in different database schemes despite differences. This is a major difference to earlier systems. We mean by those more “classical” distributed database systems, either homogeneous, e.g., SDD-1 of CCA or even heterogeneous, e.g., MULTIBASE of the same company where work of our longtime friend since, Umesh Dayal was instrumental.
Sysidor and Messidor were two earlies multidatabase systems for documentary (information retrieval) databases. About the same time, MRDSM system built at INRIA was the 1st relational one. MRDSM is due to several students we cite in the exponents we discuss soon. The system allowed for multidatabase queries to the classical, i.e., manipulating a single DB at the time only, relational DBS that was MRDS of Multics. Reference to MRDSM is still somewhere in MIT Multics archive. There was important research on multitabases systems since. As we stressed elsewhere, today virtually all major industrial relational DBSs are multidatabase. Here are some exponents showing the amount of early effort (pdf). It is a large file. So, on ADSL, allocate a minute or so for this download.
One knows well by now that the counterpart to interoperability is heterogeneity of some kind. One important facet of interoperability is to make interoperable databases with heterogeneous data models. Surprisingly, the earliest book on this subject to our knowledge was Russian by Dr. Kaliniczenko in 1983. Lost contact with since. The title and cover page are our 1st and 2nd exponents.
The next two present the December 1986 Special Issue of IEEE Computer entitled Database Architecture. The issue selected major trends, according to the Editor, Prof. Nick Roussopoulos, ACM Fellow (UMD). We were proud to be invited, by the way as the only European researchers. Our article coined the now popular concept of multidatabase interoperability. We could almost instantly see the impact of this publication on much larger technical public than our earlier publications. Perhaps since the journal was a US one, what constituted a quality stamp for many, in Europe especially. Anyhow, while we’re writing this page, quick check on Google indicates 800K+ results for the terms. The exponents are the cover page and the Table of Contents.
Our next two exponents are the title and cover page of the first major tutorial on the domain by Amar Gupta, published by IEEE in 1989. The book contains the reprints of selected major articles, for the Editor, evidently. We have contributed with Prof. H. Tirri, Finland, by the article on the value dates, already presented at my Home Page. Worth noting that couple of years later, my dear friend Henri became the head of Nokia Research in Palo Alto. The book is still around, in the time of writing Amazon sells it used for $3.90.
Next, there is another important IEEE tutorial by Bright & Hurson (eds). It appeared only four years later, in 1993. This time clearly labeled as on multidatabase systems. There are several our articles selected for this tutorial, among many major others. The book does not seem to be easily available anymore. At least Amazon is out of stock.
Next exponents are cover and title page of IMS-91, 1st Intl. Workshop on Interoperability in Multidatabase Systems. Editors were Prof. Prof. Kambayachi, Rusinkewicz & Scheth. The 1st one, sweet guy for every one knowing him, regrettably literally killed himself by too much work. Discovered in his office, after a weekend, I believe. Fortunately the two others are alive and well (2016), to the best of my knowledge. There were further IMS Workshops, but I do not have any related exponent.
Next exponent is the cover page of the 1998 book by Elmagarmid, Rusinkiewicz & Sheth. In its time, i.e. almost 20 years ago, this was probably most extensive monography on various facets of a multidatabase system design. Perhaps it is still so. Editors contributed themselves to chapters, as well as several other authors. The book presents in particular an implementation of the relational multidatabase language MSQL, by L. Suardi & M. Rusinkiewicz. Unlike our own at INRIA that one was 1st to support multidatabase transactions. See also 1993 Data Engineering Conf. The book is still on sale, for modest $137 by Elsevier. Selected pages are free at Google books.
Next exponent of several pages belongs to grey literature. It is called SIRIUS Gazette . Had a single issue for so-called Journées du Projet Pilote SIRIUS in March 1990. It mainly presents the main distributed database system (DDBS) called SIRIUS DELTA, issued from the project. It was intended as a “classical” DDBS, e.g., like SDD-1 system of CCA. In other words, it was a single although distributed) database system and not a multidatabase system as we defined it. SIRIUS DELTA code was passed to French company Intertechnique. No clue what happened to it next. Gazette presents also several other prototypes developed by the teams university teams involved in the project. This included preliminary info on first multidatabase prototypes being under construction by then. Some and yet others SIRIUS prototypes, e.g., Messidor, were later presented also at the Symposium. This exponent illustratates the magnitude of the pioneer effort involved in SIRIUS project.
Further exponents are cover pages of some important books, early Thesis on MDBS, Research Reports etc. We show again the books about our Symposium and DDSS seminars in more detail. Other memorabilia are self-explanatory. Main goal are the names of researchers or editors involved. Some results were republished as articles in journals or conferences. Some not.
Some exponents concern interactions involving other facets of our scientific activities. The title page Applications of Databases is from the proceedings of the 1st Intl. Conf. on this subject. It was hosted by the friend for many years, Prof. Tore Risch (U. of Uppsala). The idea was to focus on complete applications embedding databases (packages), breaking off somehow from more traditional conferences on core database technology. The applicative trend eventually became important, even for the majors, e.g. Oracle, Sybase Everywhere and SAP. The conference was followed up by a 2nd one, two years later in Santa Clara, CA. That one was hosted by Prof. Ketabchi, (Santa Clara U.). Amazingly, he became later the founder of such a quite large company.
Yet other exponents concern the research on scalable distributed data structures. In particular, these exponents relate to the series of international seminars on this topic.