November 15, 2006
Objects and Databases: State of the Union 2006William Cook, Moderator: I want to get to you object database people, any other comments whether the acronyms soup is going to ever end?
Derek Henninger, Progress Software: I completely agree that having a language be the access mechanism is right and an object store certainly supports that strongly, but again: If you need to get information from another source that's where it becomes impractical and you do need a standard approach for doing that in order to interact with other solutions. I would love to see a standard. Personally after following it for 10 years I've gotten tired and I'm glad Patrick is working on it because I burnt out on that, good luck!
Erik Meijer, Microsoft: Because we're running out of three-letter acronyms we're introducing four-letter acronyms, now with LINQ -- so we have some more space to grow the alphabet soup.
Craig Russell, Sun Microsystems: I think that there'll be programming languages created at pretty much the same rate as we've done it so far until we've decided that there's no need for another language. Some languages will become obsolete. But I don't see any a priori decision not to create other programming languages.
William Cook, Moderator: The rate of programming language introduction is a little slower than the object relational mappers.
I have some other questions [from the audience]. This is a question for the object database people:
"If the object winter is starting to thaw, what happened or what changed to make this happen? What is different?"
There are some comments here about all these issues of transactions and concurrent access to data and multiple representations and all the things that have heard aboutthe problems which were reported were either true or not about object databases. Is there anything different now?
Robert Greene, Versant: We've kind of moved past that initial phase of how we put pretty pictures on a web page in Java. We start to think more seriously about solving our server side with serious scalability problems with that same language. With that, I think people really at large embraced object orientated programming. Smalltalk started the revolution, nearly died initially with Java in some aspects and grew again and have people thinking again truly in ways that they haven't before.
If you get a crowd of a hundred developers out of the university now and find out about how many of them know about OO, and know it reasonably well and understand the concepts, it's probably all hundred if not 95. If you went back prior to Java and did the same thing it'd probably be about 50:50 at that point, so things have changed quite a bit. So with that people again are starting to focus on how it is to solve very, very complex problems of all themselves and they're thinking in terms of modeling problem domains and creating models which represent the solution to the thing they're trying to solve. And with that they're only now coming back to the point where there is this thing in the background, which is the way that we've always represented our data and how it is that it imposes these restrictions upon to what we're trying to do with our models.
And that's I think precipitated some frustration that had been absent for a long time and so I think that the attitudes and the problems that are trying to be solved are precipitating a change in that direction to reevaluate what is it that we're doing there, that makes it so hard.
Somebody who doesn't even know how to program can go to Second Life and they can design a three-dimensional space and they're not even a software developer. And yet we haven't figured a way to distract ourselves away far enough from the way we deal with data yet to make that not painful for ourselves. That has just got to stop.
Bob Walker, Gemstone Systems: I agree entirely, that was very well said, thank you.
Patrick Linskey, BEA: I have a question for the audience: Raise your hand if you've ever written an objects persistence framework. [About 40% of the audience show their hands.] Now raise your hand if you're still using it. [About 15% show their hands.]
William Cook, Moderator: What do you conclude?
Patrick Linskey, BEA: Developers like to write object persistence frameworks. [Laughter]
Bob Walker, Gemstone Systems: Developers like to keep things complicated.
Christof Wittig, db4objects: I think a big game changer is open source. We're not forced to eat what Microsoft and Oracle spoonfeeds us anymore. [Laughter, clapping]
Erik Meijer, Microsoft: [Smiling] I'm not even going to comment on that, I'm above these things. But the thing is that I want to go back to one of the earlier remarks that was made. I think there's a big difference in the way you store your data and the way you access it. I think especially objects are a very bad way to store data because the pointers go from the inside of the objects to the outside. That means that you can never have different relationships between objects without modifying the objects.
Here I'm schizophrenic, I think there the relational model is much better or something like RDF where you can have pointers from the outside or you can kind of take read only data and then kind of connect them, put relationships between them. I don't see how to do that with objects even if you have dynamically typed language, it would be still kind of modifying the objects to form the relationships.
Craig Russell, Sun Microsystems: I like to comment on that: The form of the data that you get depends on where you are using it. In memory for small connected devices with 2 gigs, 5 gigs, 100 gigs, what are we talkingin 5-10 yearsa hundred 100 gigs what you are going to carry around with you? Your entire collection of favorite movies? The relationships that you can navigate in memory are very different from the ones you navigate if you happen to be Paramount Pictures and want every scene cross annotated, that's a much bigger problem. But if you want to just do some query on your cached information there's no point in storing it in tables. 
Previous Page |
1 Introduction
|
2 Bob Walker, Gemstone Systems
|
3 Derek Henninger, Progress Software
|
4 Robert Greene, Versant
|
5 Erik Meijer, Microsoft
|
6 Christof Wittig, db4objects
|
7 Patrick Linskey, BEA
|
8 Craig Russell, Sun Microsystems
|
9 Ten Years from Now?
|
10 Data -- It's Scary
|
11 Acronym Soup
|
12 Business and Social Issues
|
13 Back to the Social Question
|
14 Different Viewpoints
|
15 A Bigger Picture Question?
|
16 A Service-Oriented Impedance Mismatch?
|
17 Wrap Up
Next Page
|
 |
|
|||||||||||||||||||||||||||||
|
|
|
