Skip to content
Tags

ODBMS where did they go?

by Roberto V. Zicari on September 20, 2007

September 26, 2007–

If we look back at the history of ODBMS. one sees how when the first generation of ODBMS were introduced, expectations did not really meet reality, notwitstanding all the efforts and enthusiams put on it.

So one asks himself a question. What happened to the ODBMS? Where did they go?
And perhaps a more interesting and actual question is what are the realistic chances for the “new wave” of ODBMS products?

As always, part of the answers can be given by trying to understand the past.
The first wave of ODBMS failed partially because the market conditions were not ready.
In particular, object technology (in particular object oriented languages) were not as diffuse as they are now.
So the issue of “impedence mismatch” between programming languages and databases, was mainly an interesting
academic issue, but not really felt by the market.

I visualize it like a sales person who is trying hard to sell a mini sport convertable car to a family with a number of kids who is not really interested in that…

Things have changed in the meanwhile. Object Oriented Programming Languages (OOP) are widely used.
It is estimated that today we have between 3 and 4 million Java developers. Also, very important in my opinion, for the acceptance of the use of OOP, was the decision of Microsoft to develop Csharp. The Microsoft development world has changed with the introduction of the .NET Framework.

Object Modeling is no more fragmented in several different methodologies like in the early days (the first object-oriented modelling languages began to appear between mid-1970 and the late 1980s, and went up to more than 50 during the period 1989-1994. Most notably Booch, OOSE/Jacobson, OMT).
It has now found a de facto standard, UML. UML is not merely an object modeling software methodology. Many companies are incorporating UML as a standard into their development process and products, which cover disciplines such as business modeling, requirements management, analysis and design, programming, and testing.
UML being a standard has helped the acceptance of OO technology, especially in certain domains.

And relational databases are still there… They changed, yes, but they are still there.

Moreover, new marked opportunities arise. For example, with 1.2 billion cell phones in the world, mobile software development has become a lucrative industry.

So what are the *real* chances for the new generation of ODBMS?

I have decided to work on this issue and try to come up with a reasonable answer. I am planning to collect some relevant information and hopefully this will be assembled in a white paper I am planning to write for December.
In the meanwhile, if you have any input, feedback you wish to give me, there are very welcome!

Roberto V. Zicari

From → Uncategorized

4 Comments Leave one →
  1. Hi,

    I’ve got a pair of questions on your message.

    1) This phrase –
    “And relational databases are still there… They changed, yes, but they are still there.”
    – sounds like you think that possibility exists that relational DB can disappear or become unnecesarry. Is it true?

    2) The one of main trends of long DBMS history is the separation between the ones and applications. How this idea can be joined with attempts to link current OO-languages designed for application creation (like Java or Csharp) and ODBMS. I mean these OO-languages has a lot of features which are useless for DBMS and, on the other hand, DBMS need for other language features (like transaction, security, data queries ets.). For example I think that possibility to desribe some enterprice in OO-way is enougth for real _O_DBMS language. May be special OO language should be more useful for ODBMS which allows describe the modelled enterprice as set of object and which is absolutely independent from client side aplication including their languages?

  2. 1) What I mean is that Relational Database Systems were not replaced by first generation ODBMSs and they will not be replaced by the next generation of ODBMSs either.
    They have adjusted though…

    2) Matching Programming Languages and Databases has always be a difficult task.
    It is still is.
    Now the “Impedence mismatch” is more than ever a real problem felt by the market.

  3. “ODBMS where did they go?”

    They went commercial……

    I think odbms’s went from research db’s to commercial db’s to fast. They went for the money as they thought everyone would drop relational db’s.

    That is why I like open source object databases. I think they give new and student users time to play with an odbms and understand it better. Like MySql and Postgresql.

    And odbm’s are not taught at enough universities…..well not in SA…….

    And there are more architects that don’t know what an object database is…..

  4. In your 9/26/2007, you said you were going to write a white paper for December that year. Have you done it? If so, how can I get hold of it?

    Thank you in advance.

Leave a comment to Meridian Cancel reply

Note: HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS