|
Posts: 135 / Nickname: fsommers / Registered: January 19, 2002 7:24 AM
Java EE 6
November 6, 2007 2:00 PM
|
|
In this interview with Artima, Bill Shannon and Roberto Chinnici, spec leads for JSR 316, Java Platform, Enterprise Edition 6, discuss the key design considerations for the upcoming version of the enterprise Java specification.
http://www.artima.com/lejava/articles/java_ee_6.html What do you think of the design decisions Shannon and Chinnici discuss for Java EE 6? |
|
Posts: 1 / Nickname: darthvader / Registered: June 8, 2005 9:10 PM
Re: Java EE 6
November 7, 2007 1:09 AM
|
|
Good interview...I liked the pruning part. Sun has admitted better solutions have come up which can be plugged in rather than having some less efficient ones built-in and imposed on the users.
Web Profile is another thing and if JRuby can kick in good then JEE community may have something lightweight similar to LAMP. |
|
Posts: 14 / Nickname: aeoo / Registered: April 12, 2006 7:02 AM
Re: Java EE 6
November 7, 2007 11:02 AM
|
|
I don't know how to say it any other way. As I was reading the interview my BS-alarm kept going off. The interview seems so fake and PR-like. It reads like a pure propaganda/spin job.
What a waste of my precious time. |
|
Posts: 135 / Nickname: fsommers / Registered: January 19, 2002 7:24 AM
Re: Java EE 6
November 7, 2007 1:53 PM
|
|
> I don't know how to say it any other way. As I was
> reading the interview my BS-alarm kept going off. The > interview seems so fake and PR-like. It reads like a pure > propaganda/spin job. > > What a waste of my precious time. We'd like to think that Artima is a forum where you can bring up specific points you disagree on. The more specific your points are, the more likely you will get a response from the interviewees themselves. Please mention where you think the interview is "PR." One thing that may not have come out in the interview clearly is that Java EE 6 is an early-stage JSR. That means the expert group has a good idea of where they want to go (which is what Bill Shannon and Roberto Chinnici explained here), but they don't have the specifics developed yet. |
|
Posts: 409 / Nickname: bv / Registered: January 17, 2002 4:28 PM
Re: Java EE 6
November 7, 2007 0:50 PM
|
|
> I don't know how to say it any other way. As I was
> reading the interview my BS-alarm kept going off. The > interview seems so fake and PR-like. It reads like a pure > propaganda/spin job. > > What a waste of my precious time. The last thing we want to do is waste anyone's time. Can you elaborate on what made it come across to you as fake, because it wasn't fake. I thought Frank asked good questions. Was there something in particular that rubbed you the wrong way? |
|
Posts: 14 / Nickname: aeoo / Registered: April 12, 2006 7:02 AM
Re: Java EE 6
November 7, 2007 2:10 PM
|
|
> > I don't know how to say it any other way. As I was
> > reading the interview my BS-alarm kept going off. The > > interview seems so fake and PR-like. It reads like a > pure > > propaganda/spin job. > > > > What a waste of my precious time. > > The last thing we want to do is waste anyone's time. Can > you elaborate on what made it come across to you as fake, > because it wasn't fake. I thought Frank asked good > questions. Was there something in particular that rubbed > you the wrong way? There were many things here and there. The answer to the first question was too long. There are only 3 themes, so just list the three. It should be about 3-5 sentences. Instead it's 15 paragraphs long. To me that's a clear indication that an attempt to BS has been made. Answering a 5 sentence question with 15 paragraphs, is the definition of BSing. That kind of exercise is best left to English 101 students at colleges (of course that's a cynical thing to say, because ideally English classes should not be about BSing 5 pages worth of nonsense). Second question was "Why do you think those communities formed outside of Java EE in the first place, and how do you plan to address their initial concerns about Java EE in version 6?" Seems like a leading question, but it is fair to my mind, because it portrays accurately what happened, namely, that there were indeed problems with Java EE, and that the community has simply wondered off in their own direction due to dissatisfaction with Java EE "you're a sissy unless you use all Java EE technologies" BS. Ok, so how does Bill answer it? Fasten your safety belt: "One is that we have been so incredibly successful with Java EE, and that success both enabled the Java EE community and inspired people to consider competitors to that technology." Seriously? So, there are lightweight and actually functional alternatives to many heavy weight and downright dysfunctional Java EE technologies because Java EE is so, sooooo, soooo incredibly successful. Great time to pet yourself on the back without any reservation. Java has succeeded in large part despite Sun's efforts to destroy it. Think Eclipse. Think SWT. Think Jikes (good times). Think outside-the-box approaches taken by all the good teams/projects. Think of all the great stuff (and not so great) in Jakarta, Codehaus and independent projects. Sun has been dragged kicking and screaming against their will to start considering making JVM a first-class environment for dynamic languages. The list of innovations that occurred outside of Sun and that later were appropriated by Sun is big. I have to say that Sun did a pretty good job on Java JVM itself. Sun did many good things with Java 2 SE. But when it comes to Java EE, it's been pure hype and disappointment. For a recent example, just think of the pain that JSF causes. JSF has been analyzed and criticized to death in other places, so I won't repeat anything here. Of course Bill later throws a little crumb to the community by saying that the community is not totally worthless or something like that. |
|
Posts: 128 / Nickname: watson / Registered: September 7, 2005 3:37 AM
Re: Java EE 6
November 8, 2007 7:10 AM
|
|
> I have to say that Sun did a pretty good job on Java JVM
> itself. Sun did many good things with Java 2 SE. But > when it comes to Java EE, it's been pure hype and > disappointment. Well, I would say the Servlet spec is largely a success. Creating a servlet and deploying it in one of a fairly large number of servlet containers is dead simple. It's also been remarkably stable. That part of Java EE is good, IMO. It's EJB that was never a good solution for most everyone (I'll leave open the possibility that it's perfect for someone.) You could take the idea that J2EE's success spawned the other solutions, and make it realistic. J2EE was one of the biggest marketing successes of all time. Technically, not so much. Once everyone had already bought into it, they then realized that it wasn't all it was cracked up to be and needed a Java-exit strategy and this created the opportunity for the alternatives to step in. |
|
Posts: 135 / Nickname: fsommers / Registered: January 19, 2002 7:24 AM
Re: Java EE 6
November 8, 2007 8:09 AM
|
|
> You could take the idea that J2EE's success spawned the
> other solutions, and make it realistic. J2EE was one of > the biggest marketing successes of all time. > Technically, not so much. Once everyone had already > y bought into it, they then realized that it wasn't all it > was cracked up to be and needed a Java-exit strategy and > this created the opportunity for the alternatives to step > in. One of the takeaways from this all is that one should not just adopt something because it's a "standard" of some sort. My understanding is that especially larger businesses tend to favor "standard" solutions over non-standard ones. This may give a CIO or CTO the comfort of knowing that, to paraphrase, no one ever got fired for buying into a JCP standard, but that may also not be the most effective solution. I think businesses are much more open to that notion now, after the initial EJB experience. Spring, Hibernate, and many other solutions became de facto standards, partly because, as you describe, these formed an "exit strategy." |
|
Posts: 28 / Nickname: billpyne / Registered: January 29, 2007 4:12 AM
Re: Java EE 6
November 8, 2007 9:17 AM
|
|
My comments are based on the one J2EE project I've worked on. Admittedly, I have limited experience and the project occurred when iPlanet was still a contender in the app. server market - maybe 5 years ago.
> Well, I would say the Servlet spec is largely a success. > Creating a servlet and deploying it in one of a fairly > y large number of servlet containers is dead simple. It's > also been remarkably stable. That part of Java EE is > good, IMO. Servlets were easy. No problems there. JSP's stunk mostly because step debugging was non-existent. Furthermore, the app. server would get confused frequently if you made a small change to a JSP - say added in an extra "display" statement - and didn't do a full undeploy/deploy cycle. (Redeploy frequently didn't work for some reason.) Frequently, we had to mixin JavaScript to give the users the look-and-feel they were used to from pc-side technologies. > It's EJB that was never a good solution for most everyone > (I'll leave open the possibility that it's perfect for > someone.) The largest problem I saw with EJB's was the almost religious devotion to statelessness that our architect had. The module I was responsible for had a couple dozen screens that the users wanted in tabs. Frequently, they navigated from tab A to tab G - to check a piece of information - and then navigated back again to finish filling out/committing a form. Having to requery tab A's information each time caused slowdowns in user performance, not to mention keeping the state of form A during navigation caused lots of extra work since the server couldn't help. Unfortunately, I didn't get a chance to use stateful EJB's so I don't know their pitfalls. Judging by the writeup I've seen in various books, stateful EJB's would have helped my situation tremendously. So, what I'm saying is that I'm not sure, based on my experience, that the problems we had were more related to bad technology or bad choices on our part. |
|
Posts: 28 / Nickname: billpyne / Registered: January 29, 2007 4:12 AM
Re: Java EE 6
November 8, 2007 9:27 AM
|
|
> Having to requery tab A's
> information each time caused slowdowns in user > performance, not to mention keeping the state of form A > during navigation caused lots of extra work since the > server couldn't help. After writing the comment above, I realized it was going to confuse people. To clarify, we went from requerying info. each navigation to tab A to passing the state of tab A back and forth through HTTP, which I disliked even more. Basically, the network became our state keeper. Not to mention the extra garbage code I had to put into the EJB to figure out if the request contained state or not. |
|
Posts: 128 / Nickname: watson / Registered: September 7, 2005 3:37 AM
Re: Java EE 6
November 8, 2007 2:43 PM
|
|
> Servlets were easy. No problems there.
> > JSP's stunk mostly... I just mean servlets. For a long while servlets were largely just a small piece of getting a webpage online. With the growing popularity of web services, that's not always the case. > The largest problem I saw with EJB's was the almost > religious devotion to statelessness that our architect > had. One of the biggest problems with EJBs in general was that it came with a religion. And, IMO, that religion forwarded some really crappy ideas. I have the book "EJB Design Patterns" but it doesn't see the light of day anymore. It's largely worthless AFAIAC. I remember being overwhelmed by all of the complex and painfully tedious things that doing EJB entailed. Everyone was sure that it was the 'right' way. I started to doubt EJB when we spent a ton of hours trying to work around the limitations of EJB and when I finally understood the big picture, I lost all faith in it. I digress but I guess my point is that your experience is not unique. A lot of the assumed requirements like statelessness were either going to provide benefits that often never materialized or were required in order to remain in the standard and to be vendor neutral which was also meant to provide benefits that were largely unrealized. |
|
Posts: 28 / Nickname: billpyne / Registered: January 29, 2007 4:12 AM
Re: Java EE 6
November 9, 2007 6:35 AM
|
|
> I just mean servlets. For a long while servlets were
> largely just a small piece of getting a webpage online. > With the growing popularity of web services, that's not > t always the case. As well as I can recall, we used a single servlet as a controller in an MVC architecture. I remember thinking that they could probably be used more and it would remove some of the pain we went going through. > I digress but I guess my point is that your experience is > not unique. A lot of the assumed requirements like > statelessness were either going to provide benefits that > often never materialized or were required in order to > remain in the standard and to be vendor neutral which was > also meant to provide benefits that were largely > unrealized. With hindsight, I'm trying not to "throw the baby out with the bath water". EJB's were fine IMO. We used them in a stupid way and drank the Kool Aid Jim Jones-like J2EE evangelists were selling: "stateless only because you never know if you're going to have a bazillion people using the application simultaneously". (Our application was so specialized and couldn't be used outside of our company so on a good day 200 concurrent users would be enormous.) I'm all for Sun looking at J2EE/JEE and trying to figure out what could be done better. That's the nature of good developers. However, was EJB just fine but we needed better information and education about their use? Not to remove all responsibility from Sun, were they the victims of their own hype? |