Spring vs. ATG Nucleus

[stueccles]

thats an interesting concept. Spring POJOs and javaspaces, could be very cool.

But ATGs concept of session scoped object was "one per user session". how will this not be tied to the web tier. guess you could pass the session id on the thread and use that to determine the invocation lookup ......

[Rod Johnson]

Sure, the web session implemention is tied to the web tier, but the concept of a ScopedTargetSource is not. There's a resolution strategy interface behind it that can be implemented backed by a request or session.

[stueccles]

yeah i get that. the possibilities are quite powerful. i guess the ThreadLocalTargetSource is already sort of an example of a ScopedTargetSource

just considering how you would implement a TargetSource that understands the concept of a web request or web session, Either "backed" by the request object or session which im guessing would require some kind of ThreadLocal setting variable indicating which Request or Session object belongs to this Thread. or by having a cached lookup of object invocations to sessionid/requestid and passing that in the thread. Either way its going to require the dispatcher servlet or an interceptor accessing the TargetSource to set the request/session on a ThreadLocal variable.

but maybe you can think of a better way :o) Either i look forward to seeing what you come up with.

[Rod Johnson]

Stuart

which im guessing would require some kind of ThreadLocal setting variable indicating which Request or Session object belongs to this Thread

Yes, this is how it's done. It uses a Filter to bind the request, so will work in Struts, JSF et al as well as Spring MVC.

Rgds
Rod

[adepue]

... And because the implementation isn't tied to the web layer, there are lots of interesting other possibilities, such as a per-transaction scope, or a scope transparently backed by a JavaSpace, with replication support.

As a side note here, I just want to say that JavaSpaces are awsome, and I've never understood why they haven't taken off more in environments needing clustering and scalability. I suppose many end users who don't need the more exotic features of JavaSpaces can settle just fine for some JMS implementation (such as the incredible ActiveMQ) - JMS implementations tend to be more mature as well, since they are more popular - but still, JavaSpaces is just... awsome.

- Andy

[Rod Johnson]

Quite a few Interface21 clients (especially in banking) are interested in JavaSpaces, so I've prototyped some integration. The first phase of this was synchronous remoting uses a write/take instead of a conventional remote method call. Completely transparent, using the same AOP approach as other Spring remoting. There are also lots of other possibilities. You could have transparent clusterable stateful objects using a scoping integration that should be quite simple.

One of the leading JavaSpaces vendors is also getting very interested in Spring, so we should hopefully see something happening there soon.

[marquez]

Hi all,
I don't know if this is the correct forum for my request, but this is the only Spring topic where I found Atg Dynamo mentioned.
I've used Atg Dynamo for 4 years and now I'd like migrate forward a J2EE standard architecture. I never used Spring but reading Spring Documentation it seems the J2EE framework more close to Dynamo/Nucleus idea. I think the first step in my migration should be write a J2EE Application using Spring and running on DAS 6.3.
What do the Srping gurus think about this Idea?
Have you evere ran Spring on DAS? It is theoretically possilble?

Thanks in advance
Regards
Andrea

[Rod Johnson]

It is possible to use Spring on any server running on Java 1.3 and above and complying with J2EE 1.3 APIs and above (if you want to use Spring MVC or J2EE features).

[paul.barry]

Where I work we have a few apps that we have developed since I have been here, which use Spring and other standard or open source technologies, and we also have a website that is in production currently running on ATG Dynamo. We are planning to migrate away from ATG Dynamo this fall, for a few reasons.

The biggest reason is that doing test-driven development with ATG seems to be impossible, or at least very difficult, as far as I can tell. Spring is obviously the opposite of that and we are hoping to make test-driven development a big part of our development process.

The second reason is that in my experience, ATG support is not very helpful. There are no books about ATG and no good communities (forums,mailing-lists,etc.). Learning best practices is much harder with a proprietary, commercial product like ATG. I have found that projects like Spring and iBATIS have great online documentation and great online communties that have people who know what they are talking about and are willing to take time out of their day to help you out. Bottom line, it is much easier to get an answer to a Spring related question than it is and ATG related question. In the commercial vendor space, I think ATG is far better than BEA or IBM and it is too bad that they never got a larger mindshare, because maybe there would be more helpful information (books,articles,blogs,forums,mailing lists,etc.) available.

Another reason is that ATG is expensive and some of our projects have the budget to support it, others don't. We are a small development staff (3 full time programmers) and we have to support multiple applications. We spend enough time reading Rod's books, listening to Rod's presentations and learning the best practices for the Spring framework and other related technologies like iBATIS, Hibernate, Velocity, etc. I want the lessons learned from one project to carry over to the other projects. With ATG, you have to learn how to do things their way, then for other projects switch over to the Spring way.

[hhoang]

The second reason is that in my experience, ATG support is not very helpful. There are no books about ATG and no good communities (forums,mailing-lists,etc.).

This is what happens 4 years ago when I commented on the TSS
http://theserverside.com/reviews/thr...thread_id=8096. I wish ATG was paying attention back then and release Nucleus open source because Springframework wasn't available yet. It's too late now.