EJB 3.0 simplifies enterprise bean types

G

gk

"EJB 3.0 simplifies enterprise bean types"

what is the *possible* meaning to this wording ? There is not much
relevant context I can give here.

Does that mean it does not have remote and home interface ? or there
is some more meaning attached to this ?

comments please.
 
A

Arne Vajhøj

"EJB 3.0 simplifies enterprise bean types"

what is the *possible* meaning to this wording ? There is not much
relevant context I can give here.

Does that mean it does not have remote and home interface ? or there
is some more meaning attached to this ?

It is a very imprecise statement, but I am guessing that they mean:
* annotations instead of XML
* simpler model for interfaces and implementation class

In my opinion the "simplification" is 80% hype and 20% reality,
but ...

Arne
 
L

Lew

That is what it means, from a programmer's perspective. It also means that
there's a lot less boilerplate code in your EJBs.

It doesn't mean that EJBs are now automatically the right thing to use in any
particular application.
It is a very imprecise statement, but I am guessing that they mean:
* annotations instead of XML
* simpler model for interfaces and implementation class

In my opinion the "simplification" is 80% hype and 20% reality,
but ...

The simplification is plenty real, just limited in scope. The coding model is
much simpler, and eliminates the crazy disconnection between EJBs and the
interfaces they implement. Deployment is not simplified.
 
A

Arne Vajhøj

The simplification is plenty real, just limited in scope. The coding
model is much simpler, and eliminates the crazy disconnection between
EJBs and the interfaces they implement. Deployment is not simplified.

It is simpler, but how much does it really impact the big picture.
I would be surprised if that simplification reduced development
hours with more than 0.5%.

Arne
 
A

Arved Sandstrom

Arne said:
It is simpler, but how much does it really impact the big picture.
I would be surprised if that simplification reduced development
hours with more than 0.5%.

Arne

Depends on the size and composition of the project, but you won't be far
off. Thing is too, even with EJB 1.x once you figured out the techniques for
one class of beans - CMP entity beans, BMP entity beans, SFSBs, SLSBs, how
to set them up and code them and call them - it was all boilerplate after,
and didn't add that much to the effort. For any project that has hundreds
(or more) EJBs of various types I don't suppose the simplifications matter
all that much. I'll agree with Lew that the coding model is indisputably
cleaner, but I doubt we're saving much time because of it.

AHS
 
L

Lew

Arved said:
Depends on the size and composition of the project, but you won't be far
off. Thing is too, even with EJB 1.x once you figured out the techniques for
one class of beans - CMP entity beans, BMP entity beans, SFSBs, SLSBs, how
to set them up and code them and call them - it was all boilerplate after,
and didn't add that much to the effort. For any project that has hundreds
(or more) EJBs of various types I don't suppose the simplifications matter
all that much. I'll agree with Lew that the coding model is indisputably
cleaner, but I doubt we're saving much time because of it.

It's been bugging me since EJBs came out - when is it worthwhile to use 'em?

Many developers I knew were gun-shy of EJBs, having used them. I've used them
on jobs - unwieldy sometimes, but always a few folks in the shop understand
them well. Heck, I've written and debugged them, too, but the rationale for
their existence never seemed much beyond, "The architect said to put 'em
here." Excuse me, "The Architect said ..."

One alternative is POJOs; repeat for each web app. It's not too hard to write
the same component many times (copy-and-paste helps, natch). True, there are
fragilities in the build-from-common-skeleton approach. Are they worse than
the difficulties with EJBs? Holistically, you must consider both coding and
operational effort.

Maybe it's the way people used them that's soured me. Maybe I just haven't
seen well-rationalized EJBs in practice.

I'm playing with Glassfish now, and portal, seeking the Renaissance ideal of
rapid development AND deployment. Facility with the tools should yield
insight into their proper niche.

One place I see the call for EJBs is Transaction World. and the work world
abounds with demands for multiple access points to common application
services. (Example: Separate GUI, custom XML and SOAP web service endpoint
access to an online shopping site.)

The mentality of the "new" EJB appeals to the way I think. Instead of
thinking about "Home" and "Remote" interfaces and keeping them straight from
the Java keyword sense, I'm thinking about business logic and message flow
with hooks into the enterprise framework. It's more like aspect-oriented
programming - the EJB-ness offers hatchways into shared context.

It is possible to deploy EJBs now, and JPA stuff, with relatively light
reliance on orthogonal XML deployment descriptors. Too many deployment
environments I've seen have put the "ugh" in "spaghetti". At least with the
3.0 way that is not required.

What is the true power of EJBs?
What is the right way to use them?
 
A

Arne Vajhøj

It's been bugging me since EJBs came out - when is it worthwhile to use
'em?

Many developers I knew were gun-shy of EJBs, having used them. I've used
them on jobs - unwieldy sometimes, but always a few folks in the shop
understand them well. Heck, I've written and debugged them, too, but the
rationale for their existence never seemed much beyond, "The architect
said to put 'em here." Excuse me, "The Architect said ..."

One alternative is POJOs; repeat for each web app. It's not too hard to
write the same component many times (copy-and-paste helps, natch). True,
there are fragilities in the build-from-common-skeleton approach. Are
they worse than the difficulties with EJBs? Holistically, you must
consider both coding and operational effort.

Maybe it's the way people used them that's soured me. Maybe I just
haven't seen well-rationalized EJBs in practice.
One place I see the call for EJBs is Transaction World. and the work
world abounds with demands for multiple access points to common
application services. (Example: Separate GUI, custom XML and SOAP web
service endpoint access to an online shopping site.)

The mentality of the "new" EJB appeals to the way I think. Instead of
thinking about "Home" and "Remote" interfaces and keeping them straight
from the Java keyword sense, I'm thinking about business logic and
message flow with hooks into the enterprise framework. It's more like
aspect-oriented programming - the EJB-ness offers hatchways into shared
context.
What is the true power of EJBs?
What is the right way to use them?

If you have some business logic where you can use one or more of:
- container managed transactions
- container managed thread pool
- container managed security
- strong rules about what Java features are not allowed
then using EJB's is simpler than creating custom solutions
achieving the same.

SLSB's and MDB's are quite nice. And even SFSB's can be useful.

(EB's were a great solution for a non-existing problem)

10 years ago they looked very complex, but today they look very simple.

A little bit because EJB 3.x simplified some things and a lot because
other frameworks has become much more complex.
> It is possible to deploy EJBs now, and JPA stuff, with relatively light
> reliance on orthogonal XML deployment descriptors. Too many deployment
> environments I've seen have put the "ugh" in "spaghetti". At least with
> the 3.0 way that is not required.

I have never seen the big benefits of having deployment information
spread out in dozens of source files instead of a central deployment
descriptor.

Arne
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

Forum statistics

Threads
473,769
Messages
2,569,577
Members
45,054
Latest member
LucyCarper

Latest Threads

Top