Developers using Lotus technology have been hearing the J2EE pitch from IBM for several years now. IBM has made it clear it would like its Lotus customers to move to the Java 2 Enterprise Edition (J2EE) environment of WebSphere Application Server at some point in the future, as J2EE is a central technology for Big Blue's entire software portfolio. IBM says Lotus customers will benefit from J2EE's standards compliance, high scalability, and strong transaction processing capabilities.

But companies that have invested in Notes and Domino aren't just going to dump their old applications and data and move to J2EE, at least not in the near term. The migration will be gradual, and IT teams will have to figure out how to use the old investments in a new environment. How can Lotus customers manage the pressure to adopt J2EE, and when should they begin the initial steps of the transition?

To address those questions, Advisor spoke with Mark Dixon, a member of the Notes and Domino community who says the belief that companies have to transition to a J2EE environment immediately is unfounded. Dixon co-founded Teamstudio in 1996 and has been the company's CTO since 1998. The company offers products that cover the range of Notes/Domino development and administration tasks (source code version control, developer productivity, etc.). It has also launched a product line for Java developers, which it will expand in the future.

In a statement released by Teamstudio, Dixon counters the widespread fear that Lotus-based development is coming to an end.

"With all the hype WebSphere is getting, it's very easy to assume that everyone else's migration is well under way," he says, "and if you don't hurry, then you'll be left at a competitive disadvantage. Nothing could be further from the truth. It's a stressful time to be a Notes developer."

Advisor asked Dixon to elaborate on the Domino-to-WebSphere transition and what it means for Domino customers' future.

Some say a full transition to WebSphere is unnecessary, that Domino is still a better choice for certain situations. Do you predict a more limited transition for most companies, in which they still use Domino for key applications?

I think you have to look at it from the perspective of what's happening in the future, as well. You have to have some concerns over how long Domino is going to be actively maintained by IBM -- whatever it says. My view is that at some point in the future, a more J2EE-based solution is going to be the platform of choice for most applications. But that's a long way out.

A popular idea seems to be that smaller firms can do fine using just Domino if they don't have to support heavy transactions or high scalability -- say, if they're running a small e-commerce site.

I think there's a lot to come in the future on J2EE, because for anti-Microsoft reasons, some are focusing heavily on the Enterprise JavaBeans [EJB] side of it, which is the high-scalability, high-end, difficult-to-get-your-head-around stuff. Actually, servlets and Java Server Pages [JSPs] are not that big a deal. For those lower-end companies, I think that's the type of solution that would work for them in the Java space.

Until some get that stuff sorted out, then J2EE is unattractive. Something is going to have to give on that whole J2EE complexity issue. Many people are put off J2EE by perceived complexity, but almost all of that complexity "lives" inside the EJB part. Some are working hard to make you think EJB when you think of J2EE, and they're making the problem worse. That's why J2EE has this horrendous reputation for complexity, and that isn't sustainable. That's why you'll see more offerings such as WebSphere Express, which is just servlets and JSP [no EJB support]. J2EE will be brought down-scale to make it appropriate for smaller companies.

As IBM makes it clear that you can make the move without having to learn and deal with EJBs, do you expect more Domino developers to hop on board?

You're in a different case if you're an existing Domino developer. Domino has always been a hard sell to the "green field" site for Web development because it has such an unusual Web development model. But if you're an existing Domino shop, it's a whole different ball game: You need to justify moving away from it, rather than choosing it.

Speaking for ourselves, our whole company infrastructure is based around Domino. We have no plans to move away from that in the short or medium term. It makes no sense to us to do that.

What strategies might IBM try to give Domino and Notes companies a stronger push into the J2EE world? Do you expect product support deadlines, dropping features from the next version of Domino, etc.?

I think IBM has already shown it's thinking of using the stick more than the carrot. Remember the "Garnet" feature set [a JSP engine IBM had planned to add to Domino] that got dropped from Domino 6 -- adding JSP and servlet support natively to Domino? [That support] would have been a great way of encouraging people to explore those technologies in a framework with which they were happy. That stuff got dropped. One of the many excuses was it didn't have EJB support, so it would confuse people: You can't have J2EE without EJBs. So it was pretty strange to see IBM announce you could get WebSphere without EJB.

That would have been the nice way of encouraging people to move over from Domino to Java -- by letting them experiment at their own pace. The remaining stick is to start decreasing support for Domino.

Most Domino companies believe they'll eventually move to J2EE. Do the companies you talk to plan to leave Domino altogether, or will Domino and WebSphere co-exist?

I would view Domino as a data store continuing to exist for a long time, if not indefinitely. There's no reason why Domino can't be up there with relational databases as certain types of data for certain situations, a viable data store of choice. I'd love to see IBM take that route, to focus on the things the Notes data store does well. The reason Notes is hard to sell into a "green field" site for Web development is the Web development model is so strange. By moving to a standard J2EE server-based presentation layer and using a Domino back end, you have a model that gives you the best of both worlds.

I can see a transition where you move your presentation away from Notes clients or Domino-generated HTML towards a Java front end, sharing data with existing back ends. Maybe over time some of that data would move out of the Domino store towards a relational database, if that makes sense for it. But I'm not sure if that Domino data ever has to go away.

And the Domino Java APIs are already there to connect to a Java front end.

Right.

Lets talk about the Lotus Domino Toolkit for WebSphere Studio. IBM says the toolkit lets Domino people move to WebSphere without having to know too much about Java. Do you have experience with the toolkit? Is IBM's message accurate?

It's a very early incarnation of what they're trying to do. [Reports are that] IBM's having a pretty hard time with this thing. It's had an amazingly small number of downloads, and people have reported a number of issues with it. At this stage, I think it's more IBM trying to give you some reassurance, rather than a viable solution. I see no reason why the principle isn't sound: Accessing Domino data from JSPs is a viable thing to do.

What are some of the biggest technical challenges to the migration, of taking the entire Domino application and porting it to WebSphere?

The biggest challenge, if you're a Domino developer and you come in Monday morning and fire up WebSphere Studio Application Developer, is you've got this big white screen in front of you. You've got no obvious starting point. All the infrastructure you're used to having -- forms, views, designers -- is gone. For every service you want, you have to find some way of recreating it before you can even think about writing a line of application code. You just have a huge replacement infrastructure to build or learn.

Some people are claiming to have some tools that will convert .NSF, but I think they're just trying to get you in the door to look at their consultancy offering, where they do the conversion by hand. The two environments are so different that even if you could auto-convert the application, what you end up with would be unmaintainable.

What advice do you offer to companies planning a migration strategy?

I would be experimenting with WebSphere. It doesn't even have to be WebSphere -- start experimenting with JBoss [an open source application server] or Tomcat [an open source servlet container used in the reference implementation for Java servlet and JSP technology]. You don't even have to spend any money. Start accessing some Domino data and learn how, writing a pure J2EE application, you'd do the kind of things you're doing today. Don't assume that your goal is to take your existing collection of Notes databases and transfer them over to WebSphere. If they're working fine in Notes, you don't have anything much to do with them -- leave them as they are. This is the Web, and it's easy enough to integrate, in a portal front end, Domino data in one place and Java-generated data in another. You don't have to convert everything.

The scenario you describe suggests Domino IT teams could split into two groups: one that handles old Domino data on the back end, and one that handles the new Java stuff. Are Domino developers in danger of being locked into the role of maintaining legacy systems?

From what we're hearing from the people we speak to -- we phone hundreds of Notes developers every week -- we aren't finding many who are doing Notes and Java development. The vast majority are continuing to do Notes development, and then a minority are no longer there because their functions are being subsumed by a separate team doing the Java work. All these people who are saying, "If you're a Notes developer, now would be a good time to learn Java" -- I think that's true, just from a personal career development point of view and regardless of what IBM is doing.

To get your feet wet with Java technology, see Brian Good's Java series in Lotus Advisor magazine, which starts with "Java for Domino Developers" (http://Advisor.com/doc/09731). Pro-level subscribers can read the entire series online.