Discussions

News: Edgar Codd, database theorist, dies at 79

  1. Edgar Codd, database theorist, dies at 79 (25 messages)

    Edgar F. Codd, a mathematician and computer scientist who laid the theoretical foundation for relational databases died on Friday; he was 79. Codd wrote several papers oon relational theory while at IBM in the 60' and early 70's, but it wasn't until the late seventies that Oracle and IBM released RDBMS.

    Edgar Codd, database theorist, dies at 79.
    For some theoretical background, read: Relational databases: The untold story.

    Threaded Messages (25)

  2. I don't know much about him, but I know that when I first discovered relational databases, they were the first thing that struck me (a set theory and abstract algebra fan) as truly beautiful. It's amazing that someone could realize that relationships could be described so economically and gracefully, and come up ways that a data domain could maitain consistency (ACID, etc).

    Unfortunately, these losses will probably start happening a lot more as people who were the early pioneers of information technology start to pass away. Each one makes me realize that the systems we use now have only begun to implement their ideas.

    -A
  3. Same here, set theory and relational algebra I learned in elementary school, but they did not say who invented it and no computers.

    Than much later I learned SQL and Cood join from Celko books.

    .V
  4. R.I.P.[ Go to top ]

    SELECT * FROM "Gratitude"
    WHERE Gratitude.ForDatabaseDevelopment LIKE "Priceless"
    ORDER BY Gratitude.LastName

    Returns "Cood, Ellison"
  5. Tribute[ Go to top ]

    Just remembering those days I used to learn Coad’s RDBMS theories religiously.

    Tribute to the genius
  6. Coad?[ Go to top ]

    Just remembering those days I used to learn Coad’s RDBMS theories religiously.

    >
    > Tribute to the genius

    No insult intended, but I'm sure you meant Codd. I believe Coad is a current OO enthusiast with Togethersoft.
  7. What a great man!!![ Go to top ]

    What a pity to lost this great man!!!
  8. Tribute[ Go to top ]

    I still remember the Codd's rules for a perfect RDBMS helped in one of my Interview. Thanks to Codd and praying for his soul.

    Jiji
  9. Relations will certainly rank as one of the greatest contributions made to Computer Science and Information Technology by this man. He will be remebered as long as mankind uses IT as a tool to furthur its progress.
  10. Tribute[ Go to top ]

    Need the constraint - 'Cannot delete the genius'.

    Praying for his soul.
  11. Hi,

       I'm 42. I've been doing computers as long as there have been desktops (early 80s). My parents are in there mid-seventies, and just started using computers a few years ago, but now have wideband, father even has his own website ( http://www.unc.edu/~schaefer ). But they will die sooner than I would like.

       It strikes me that IT is reaching middle age in a way also. Our conceptual parents, like Mr. Codd, are dying. A sobbering thought for an industry that is remaking itself every 3 to 5 years, ever young.

    -Tom
  12. We truely rest on the shoulders of giants.
  13. I am wondering whether the world recognized him with any major awards like the "ACM Turing Award". Is there is no version of the Nobel Prize in Computer Science? We certainly must honor the people who changed the way we do computers. Some of the names I recall as greats are :

    Dijkstra, Tim Berners-Lee, Codd, Bresenham, Sutherland, Diffie, Bill Joy, Knuth

    This lists a few pioneers.
    Pioneers in Computer History

    Folks, want to add more information on our pioneers.
  14. The greatest tragedy[ Go to top ]

    It's tragic for the computing industry that he died.

    However the greatest tragedy is that his theories to this date are so little understood.
    One example is the J2EE comunity that doesn't seem to want to create database access utilities that are purely for relational data. They always want to support all kind's of data storage systems, particularly socalled object databases. The result is that we have no streemlined, optimized relational oriented data access.
    One example is the EJB entity beens with it's query language - very far from what it should have been had the designers consentrated on relational databases - or even understood them properly.
    JDO doesn't seem much better, although I haven't yet studied it sufficiently.

    Why?
    May be because people don't want to study and realy understand?
    May be because most people learned in school (or preschool was it?) that theory and practice are to different things? And they never heard of Aristotle or anyone else who taught differently?

    Whatever it may be it's a crisis that I can only hope someone will one day want to do something about.
    In the meantime I do not hold my breath and have to do as well as I can with the tools that are available.

    Nils Myklebust
  15. Re : The greatest tragedy[ Go to top ]

    Nils :
    What a wonderful point you made. Just what i have been thinking of for a long time.

    Relational Databases, are one of the best persistence mechanisms that we have in the business computing world. They are mature, stable and reliable...

    We have had so much talk of Object Databases and what not, but how many people actually get down to use it succesfully in production environments... ?

    J2EE, had a perfect case for going for RDBMS as a de-facto storage mechanism.. which they did not pursue... Almost, all of us infact use one RDBMS or another for persistence while developing J2EE apps... yet the level of support for RDBMS in J2EE is low. Al in the name of giving support to anything and everything.

    Also, i have always wondered, about the hype surrounding UML ! In most business apps i do, it is the database design (rdbms) we do first, all our functionality is captured very neatly there, only then we move on to doing object design etc...
    This has worked always well for us. In every project where i have seen people doing Object design first, and leaving database design as one of the last things to do... i have seen severe gap between the requirements and implementation...among other things.

    Thanks to this great soul. May his soul rest in peace.
  16. Re : The greatest tragedy[ Go to top ]

    Too bad, we've suffered a great loss.

    And I agree with the sentiments regarding the unfortunate ignorance of Codd's ideas. This is also a great loss.

    One thing I could never understand in J2EE, was the desire of many to "hide" the complexity of RDBMS's. I consider the Relational model to be simple, powerful and elegant. As a result, I tend to create applications which are more database-centric that a typical J2EE application.

    Often, I will avoid EJB's for anything more simple CRUD operations. Even before Java, (and especially during the client/server boom) there was a problem of too many developers treating an RDBMS as a big indexed file, rather than viewing data as sets. I can only hope that new programmers will learn more about the relational model. I've seen too much software that is buggy and slow because of programmers who don't understand the relational model.
  17. Also, i have always wondered, about the hype surrounding UML ! In most business apps i do, it is the database design (rdbms) we do first, all our functionality is captured very neatly there, only then we move on to doing object design etc...

    > This has worked always well for us. In every project where i have seen people doing Object design first, and leaving database design as one of the last things to do... i have seen severe gap between the requirements and implementation...among other things.
    >
    This is pretty much my experience. I am also quite certain that there are very good reasons for it

    (1) Unlike code and especially unlike diagrams(!) actual physically stored datastructures stabilize development. Even more, agreeing to a datastructure that will eventually be physically implemented communicates a certain sense of urgency to the customer.

    (2) Datamodels are inherently easier to understand then object models because they separate behaviour from data. One of the biggest problems with OO projects is that most developers did not understand what object oriented programming actually is and how it can be efficiently used.

    (3) There is no way to actually translate a single UML digram model into single and meaningful (multi-tier, of course) code. Most of the time there are multiple diagrams at work at the same time (and need to be) or there has to be a large number of metadata definitions that need to be in place. This is emphatically called "Model Driven Architecture" although I haven't seen a single case where the model actually drove the architecture.
  18. Object v. Relational[ Go to top ]

    I too have been baffled by the relative ignorance of relational theory among working software engineers. Relational databases benefit not only from their relative maturity (for which we might allow a handicap in comparing to object databases), but also from having a rigorous theoretical grounding. Chris Date (who gets short shrift in the CNET article as "a former business partner" when in fact he was more importantly an academic colleague and is one of the pioneers of the field) has written extensively about the lack of a rigorous theoretical base for object databases, but is the lone voice as far as I can tell. I picked up a fat textbook on ODBMS just recently which dwelled almost exclusively on physical implementation details and showed a woeful ignorance of relational theory. If this is what get published then there is little hope for this to change.

    Steve Pearson
  19. may his soul rest in peace[ Go to top ]

    Like many others Codd had the deep insight that helped him develop a theory so powerful yet so elegenat and simple that anyone willing to put in a little effort can understand it.
    All I can say is that thank you. you have made the world a more ordered place.

    Dev
  20. Tribute to Codd[ Go to top ]

    May his soul rest in peace
  21. Thanks for not putting patents on your ideas. If someone else had patented Codd's ideas, the IT landscape would be much different.
  22. I too remember my first job interview in my life. Cood's theory helped me to get my first job. Tribute to our fallen heores!!!
  23. Great theorist[ Go to top ]

    Edgar F. Codd's role is in fact quite unknown to most IT professionals, despite the fact that he received a Turing Award back in '81.

    You might want to read IBM's remembrance at http://www-3.ibm.com/software/data/news/efc.html

    CNET has a story as well: http://news.com.com/2100-1012-997975.html

    Last year was also sad, loosing pioneers such as Simula- and OO inventors Nygaard and Dahl. Djikstra also recently passed away.
  24. too busy...good comments[ Go to top ]

    I've been quite busy with a group of civil engineers, and have not checked in lately. Excellent, but sad comments about some
    of our late luminaries.

    We have only been at this software stuff a relatively short while, and I am reminded that bridges and roads have been engineered
    for many centuries.

    We are getting better at delivering solid software every day. Our continued debate would honour our predecessors.

    Best,

    rga
  25. Integrity is always there![ Go to top ]

    It is undeniable a loss to our IT industry and to an extend a loss to our information world that relies heavily upon the relational database. Our recognition of his master piece of work will always remain consistent and integrity!
  26. As any true genius, Edgar Codd was so far ahead of his time that many of his contemporaries in the industry failed to appreciate the breadth of his ideas for quite some time. His theories are, to this day, the bedrock of some of the most advanced and critical systems created, without which modern enterprise and information systems would have been quite impossible. One cannot imagine how different our world would be today, had the relational database and all the tertiary systems build around it not been around.

    It is sad that there isn’t as recognizable an accolade as the Nobel Prize for Computer Scientists – Codd would certainly have been one of its recipients, and the world would have shown its appreciation. He will be remembered.

    Reuben Cleetus.