Friday, February 08, 2008

New Eclipse-Based Tools Offer Developers More Choices, Migrations and Paths to IBM WebSphere

Transcript of BriefingsDirect podcast on Eclipse-based tool choices for IBM WebSphere shops.

Listen to the podcast here. Sponsor: Genuitec.


Dana Gardner: Hi, this is Dana Gardner, principal analyst at Interarbor Solutions, and you’re listening to BriefingsDirect. Today, a sponsored podcast discussion about choices developers have when facing significant changes or upgrades to deployment environments. We'll be looking at one of the largest global installed bases of application servers, the IBM WebSphere platform.

Eclipse-oriented developers and other developers will be faced with some big decisions, as their enterprise architects and operators begin to adjust to the arrival of the WebSphere Application Server 6.1. That has implications for tooling and infrastructure in general.

The platform depends largely on the Rational Application Developer (RAD), formerly known as the WebSphere Studio Application Developer. This recent release is designed to ease implementations into Services Oriented Architecture (SOA) and improve speed for Web services.

However, the new Rational toolset comes with a significant price tag and some significant adjustments. Into this changeable environment, Genuitec, the company behind the MyEclipse IDE, is offering a stepping-stone approach to help with this WebSphere environment tools transition.

The MyEclipse Blue Edition arrives on March 15, after a lengthy beta run, and may be of interest to developers and architects in WebSphere shops as they move and adjust to the WebSphere Application Server 6.1.

To help us understand this transition, the market, and the products we are joined by Maher Masri, president of Genuitec. Welcome to the show, Maher.

Maher Masri: Thank you, Dana.

Gardner: Also, James Governor, a co-founder and industry analyst at RedMonk. Welcome, James.

James Governor: Hi, Dana.

Gardner: James, let’s start with you. We’re looking at a pretty dynamic marketplace around tools. There are certainly lots of different frameworks and approaches floating around. Folks are dealing with SOA, with Software as a Service (SaaS), with agile development. They are dealing with mashups and Enterprise 2.0 issues. We’re seeing increased use of REST and SOAP. This is just a big, fluid, dynamic environment.

On the other hand, we’re also seeing some consolidation around runtimes. Organizations looking to cut cost and infrastructure and trying to bring their data centers under as few runtime environments as possible. So, we’re left with somewhat of a conundrum, and into this market IBM is introducing a major upgrade.

Maybe you could paint a picture for us of what you see from the enterprises and developers you speak to on how they deal with, on one hand, choice and, on the other hand, consolidation.

Governor: It's a great question. In this industry we can expect continuing change. If anything is certain, it's that. When we look at this marketplace, if we go back a couple of years into the late 1990s, there was a truism that you could not make money as a tools company. The only way you could really sustain a business would be connected to, and interwoven with, the application server and the deployment environment. So it's interesting that now, sometime later, we’re beginning to rethink that.

If you look at a business like Genuitec, the economics are somewhat different. The Eclipse economics, in terms of open source and the change there, where there is a code based being worked on, have meant that it's actually easier to maintain yourself as an independent and work on a specific set of problems.

In terms of your question about Web 2.0, agile development, and so on, there are an awful lot of changes going on. That does create some opportunities for the third parties. Frankly, when you look at the very largest firms, it's actually quite difficult for them to maintain the sorts of innovation that we’re seeing from some of the smaller players.

In terms of the new development environments, it might be something like the fact that we’re seeing more Ruby on Rails. P scripting languages continue to be used in the enterprise. So, supporting those is really important, and you are not always going to get that from the lead vendors.

I'll leave it up to Genuitec to pitch what they do, but one of the interesting things they did, which you certainly wouldn’t have seen from IBM, was a while back, when they bridged the Eclipse world with the NetBeans ’ Matisse GUI building application development set.

Crossing some of those boundaries and being able to deal with that complexity and work on the customer problems, it's not surprising to me that we’ve seen this decoupling, largely driven by open source. Open source is re-enabling companies to focus on one thing, rather than saying, "Okay, we've got to be end-to-end."

Gardner: So, we've got a dynamic environment. We have some amazing uptake in Eclipse over the past several years becoming a dominant job oriented IDE. We have WebSphere as the dominant deployment platform.

As you pointed out, the economics around tools have shifted dramatically. It seems that the value add is not so much in the IDE now, but in building bridges across environments, making framework choices easier for developers, and finding ways of mitigating some of these complexity issues, when it comes to the transition on the platform side.

Let’s go to Maher. Tell me a little bit about why Eclipse has been so successful, and do you agree that it's the value add to the IDE where things are at right now?

Masri: Let me echo James’ point regarding the tools environment, and software companies not being able to make money at that. I think that was based on some perceived notion that people refuse to pay money for software. In fact, what we've found is that people don’t mind paying for value, and perceived value, when it’s provided at their own convenience and at their own price point.

That’s why we set the price for the MyEclipse Enterprise Workbench at such a low point that it could be purchased anywhere in the world without a series of internal financial company decisions, or even a heartbreaking personal decision.

Although the product was just the JSP editor when it was first launched, today it's a fully integrated development environment that rivals any Tier 1 product. It's that continuity of adding value continually with every release, multiple releases within the same year, to make sure that, a) we listen to our customer base, and b) they get the value that they perceive they need to compensate for the cost that we charge them.

Eclipse obviously has become the default standard for the development environment and for building tools on top of it. I don’t think you need to go very far to find the numbers that support those kinds of claims, and those numbers continue to increase on a year-to-year basis around the globe.

When it started, it started not as a one-company project, but a true consortium model, a foundation that includes companies that compete against each other and companies in different spaces, growing in the number of projects and trying to maintain a level of quality that people can build upon to provide software on top of it from a tools standpoint.

A lot of people forget that Eclipse is not just a tools platform. It's actually an application framework. So it could be, as we describe it internally, a floor wax and a dessert topping.

The ability for it to become that mother board for applications in the future makes it possible for it to move above and beyond a tools platform into what a lot of companies already use it for -- a runtime equation.

The next Ganymede 3.4 and the 4.0 extension of Eclipse is pushing it in exactly that direction. The OSGi adoption is making a lot of people reconsider their thought in terms of, "What application do I write for productivity applications internally, for tools that I provide to my internal and external customers, for which client implementations?"

It's forcing quite a bit of rethinking in terms of the traditional client/server models, or the Web-only application model, because of productivity requirements and so on.

IBM was the company that led the way for all of the IBM WebSphere implementation and many of their internal implementations. A lot of technologies are now based on Eclipse and based on Eclipse runtime.

Gardner: So, we have this big bear, Eclipse, in the market and we have this big bear, WebSphere, in the market. Why is there a need for someone like you to come in between and help developers?

Masri: The story that we hear internally from our own customers is pretty consistent, and it starts with the following. "We love you guys. You provide great values, great features, great support, except I cannot use you beyond a certain point." Companies for whatever internal reasons, from a vendor standpoint, are making the choices today to move forward with WebSphere 6.1, and that’s really the story we keep hearing.

"I am moving into 6.1, and the reason for that is I am re-implementing or have a revival internally for Web services, SOA, Rich-net applications, and data persistence requirements that are evolving out of the evolution of the technology in the broader space, and specifically as implemented into the new technology for 6.1."

Gardner: They need to modernize it.

Masri: But their challenge is similar. Every one of them tells us exactly the same story. "I cannot use your Web service implementation because, a) I have to use this web services within WebSphere or I lose support, and b) I have invested quite a bit of money in my previous tools like WebSphere Application Developer (WSAD), and that is no longer supported now.

"I have to transition into, not only a runtime requirement, but also a tools requirement." With that comes a very nice price tag that not only requires them to retool their development and their engineers, but also reinvest into that technology.

But the killer for almost all of them is, "I have to start from scratch, in the sense that every project that I have created historically, my legacy model. I can no longer support that because of the different project model that’s inside."

For example, Rational 7.0 is only one of the few versions of WebSphere that supports 6.1 and supports all of the standards for Web services, for AJAX support, for persistence requirements that they need to modernize. They have to implement it, but cannot take, for example, an existing WSAD project, import it into Rational 7.0, and continue development. They pretty much start from scratch.

Gardner: Let’s go to James for a moment. James, you’re familiar with the IBM stack and their road map. Why are they doing this? It seems to me that there is an application lifecycle management (ALM) set of benefits that the Rational toolset and platform bring that IBM is trying to encourage people to take advantage of. It does require transition, but they have a larger goal in mind. Perhaps we should address this ALM, or do you have other thoughts about this transition?

Governor: From an IBM perspective, it’s a classic case of kind of running ahead of the stack. If you see the commoditization further down the stack, you want to move on up. So IBM looks at the application developer role and the application development function and thinks to itself, "Hang on a second. We really need to be moving up in terms of the value, so we can charge a fair amount of money for our software," or what they see is a fair amount of money.

From an IBM standpoint, I think they really looked at players such as Genuitec, looked at where Eclipse was going, and they thought, "Wait a second. We really do need to be moving forward with this notion of software development."

If you talk to a lot of developers, they don’t really think of the world that way, but many of their managers do. So, the idea of moving to situation where there is better integration of the different datasets, where you've got one repository of metadata moving forward with that kind of stuff, that’s certainly the approach they are taking.

The idea is you've got "auditability," as you build applications. You’re going from a classic distributed development, but you’re doing a better job of centralizing, managing, and maintaining all the data that’s associated with that.

The fact that IBM is making that change is indicative of the fact that when they look at the market more broadly, they think to themselves, "Well, where is our margin coming from?"

IBM’s strategy is very much to look at business process as opposed to the focus on just a technical innovation. That certainly explains some of the change that's being made. They want to drive an inflection point. They can't afford to see orders-of-magnitude cheaper software doing the same thing that their products do.

Gardner: As we mentioned earlier, there are so many complexities involved in decision making now, different approaches to creating services, that the operators and the vice presidents of engineering are saying, “Wow, we need to manage this complexity.”

They are looking for life cycle approaches, ways of bridging design time and runtime. IBM is addressing some of these needs, but, as you point out, developers are often saying, "Hey, I just want my tool. I want to stick with what I know." So we’re left with a little bit of a disconnect.

I’m assuming, Maher, that this is where you’re stepping in and saying, "Aha, perhaps we can let the developers have it their way for a time to mitigate the pain of the transition, at the same time recognizing that these vice presidents of engineering and development are going to need to look at a much more holistic life-cycle approach. So, perhaps we can play a role in satisfying both." Am I reading too much into that?

Masri: No. We understand internally that different technologies have different adoption life cycle behind them. ALM is no different. It’s going to take a number of years for it to become the standard throughout the industry, and it is the right direction that almost every company is going to have to face at some time in the future.

The challenge for everybody, us and IBM, is the bottom-up sale process, to provide the tools and the capabilities for companies to embrace, for people to embrace those technologies, and, at the same time, putting the infrastructure in place for managers to be able to continue to manage projects into success.

Our decision is very simple. We looked at the market. Our customers looked back at us and basically gave us the same input. If you provide us this delta of functionalities, specifically speaking, if you’re able to make my life a little easier in terms of importing projects that exist inside of WebSphere Application Developer into your tool environment, if you can support the web services standard that’s provided by WebSphere.

If you can integrate better with ClearCase from a code management standpoint, and if you could provide a richer deployment model into WebSphere so my developers could feel as if they’re deploying it from within the IBM toolset, I don’t have the need to move outside of your toolset. I can continue to deploy, develop and run all my applications from a developer's standpoint, not from an administrator's.

Obviously if you are an administrator and have one to three people within the company that maintain a runtime version of WebSphere, you will need specific tools for that. We’re not targeting those one to three people. We’re targeting the 10 to 500 developers internally that need to build those applications. That’s really where Blue is coming from.

Governor: Maher, can you be a little bit more specific about it. You just used the top-down bottom-up or top-down in terms of your argument. Can you talk a little bit more to sort of that and your sales staff?

Certainly, from RedMonk’s standpoint, we do tend to be more aligned with the bottom-up, just in terms of our customer and community base. But, in terms of what you’re seeing and saying, how is what you do different from IBM? I didn’t quite get that from your last comments.

Masri: I'll give you a very simple example. Just take the experience of a developer installing MyEclipse or installing RAD from ground zero. MyEclipse, you can install in a two-megabyte root install. It installs a 600-megabyte version on your desktop that contains all the tools. You no longer need to buy additional tools from somewhere else. If you need to do UML development, if you need to do UI design, all that is included as one bundle within MyEclipse.

If you install RAD, you need a multi-DVD, six or seven gigabytes, I understand, in order just to begin the installation. The configuration is a nightmare. Everyone is telling us that it's a very difficult configuration process just get started.

MyEclipse is part of a very rich, simple profile that a user can download directly through the MyEclipse site or through our managed application environment inside of Pulse. You can be up and running with tools, with runtime configurations, and with examples, literally within minutes, as opposed to within hours or days beyond that.

On the issue of simplicity, the feedback that we keep getting is that our response level in terms of request for features, request for innovations, request in the technologies, we can deliver within months, as opposed to years or multi-months, when looking at the competition. All of that becomes internalized from the developer standpoint into, "I like this better, if it can bridge that gap that I now have to use this technology, in order to satisfy my business requirements."

Gardner: Perhaps another way of asking a similar question is: you are in beta now. You’re going to be coming out on March 15 with MyEclipse Blue Edition. What's the difference between MyEclipse and MyEclipse Blue Edition?

Masri: Excellent point. MyEclipse Blue Edition is inclusive of all MyEclipse professional features. It’s roughly on the order of 1,000 to 1,500 features above and beyond what the Eclipse platform provides, as well as the highly targeted functionalities that I mentioned. It can import and manage an existing project that you had previously inside WebSphere application developer and can develop to the Web services SOA standards that are specified into the WebSphere runtime.

It has much better integration into IBM code management, ClearCase technology, and almost an identical implementation of what you possibly could see inside Rational for deployment model and the ability to debug an existing project or a new project into the runtime environment.

Gardner: Developers, of course, are hard to come by in a lot of regions around the globe. There’s a lot of competition. Organizations like to keep their developers happy and productive. At the same time, they need to deal with some of the complexity issues of moving to SOA. If they're WebSphere shops, they know that they are going to be tied into that for some period of time. It does sound like you are trying to give both of these parties something to be a little bit cheery about.

Governor: The one of the things that I think is important about open source and understanding open source in the enterprise, but also more broadly. Sometimes you think about open source as a personal trainer for proprietary software companies. You've got these fat, flabby toys and they need to get a life. They need to get on the treadmill. They need to get thinner and more agile. They need to get more effective. Frankly, it was ever thus with IBM. IBM is a pretty big beast.

Let me go back to the old mainframe times to think about Amdahl as a third party. When the IBM salesperson came in, you always made sure you had an Amdahl mug on the desk, right in front of the salesperson. Obviously, we’re a few years on now, but that dynamic remains important. As much as organizations balance BEA WebLogic and WebSphere against one another, or WebLogic and JBoss Application Server against one another, you would also want a balance in your toolsets.

One interesting thing here is that because you've got the specificity around WebSphere, and the sort of value prop the third party is putting forward, you're able to start that balance, that conversation to drive innovation, to drive price down. That’s one of the really useful things that Eclipse has enabled and delivered in the marketplace. It helps to keep some of the bigger vendors honest.

Gardner: So, the need to support heterogeneity is going to remain in both tools and runtime, but we’re also facing the time when heterogeneity isn’t going to include hybrid approaches to deployment. And so, we’re seeing more people interested, particularly if they are ISVs or perhaps small- to medium-size businesses in taking advantage of some of these cloud-computing options. I'm thinking of course of Amazon and some others. Tell us, Maher, how this choice in tool and heterogeneity plays into some of these hybrid approaches of deployment in a cloud of some sort.

Masri: Let me expand on James’ point and then I’ll add to it. I just want to make sure that we’re not trying to present MyEclipse Blue as if we are trying to compete with IBM, which is really could be easily perceived there. What we see is an under-served market and people that are trying to make the decision, but cannot afford to make that decision.

There are companies that are always going to be a pure IBM shop and no one is going to be able to change their mind. The ability to provide choice is very important for those that need to make that decisions going forward, but they need some form of affordability to make that decision possible. I believe we provide that choice in spades in our current pricing model and our ability to continue to support without the additional premium above that.

Going forward, I fully agree with you that the hybrid model is very interesting, and we see it in the way that companies come back to us with very specific feedback on either MyEclipse or our Pulse product. There's quite a bit of confusion out there, in terms of how Web 2.0, Rich Internet Application (RIA), and Rich Client Application are designed and geared to provide and all the underlying technology to support that in terms of runtime.

There seems to be a dichotomy. I could go in the Web 2.0 world and provide a very rich, all Web enabled, all Web centric technologies for my end-users because I need to control my environment. The other side of that is the rich client application, where I have to have some form of a rich client implementation with full productivity applications for certain people, and I have to divorce the two because there is no way I can either rely on the Web or rely on the technologies or rely on anything else.

Everyone that we’ve talked to so far has a problem with that model. They have to have some form of very strong, rich implementation of not necessarily a very fat client, but some form of a client on the end-user’s desktop. They need to be able to control that, whether you are using very specific implementation of Web Services, talking to somebody else’s Web services, need to use a very specific persistent architecture, or have to integrate with other specific architectures. It gets very dicey very quickly.

That’s really where we saw the future of the market. This is probably not the right time to talk about this specifically, since the topic is Blue, but that’s why we also moved into the managed-application space and into our other product line called Pulse. This is for end-users who are using Eclipse-based technology right now, and in the future far more than that. They'll be able to assemble, share, deploy and manage a stack of applications, regardless of where those applications reside and regardless of the form of technology itself.

Take, for example, a rich-client runtime of Eclipse running on someone’s desktop. All of a sudden, you have a version of software that’s you can deploy and manage, but it already has an interface into a browser. You can provide other Web 2.0 and RIA models, as well as other rich Internet technology, such as a Flex and Flash. These technologies are merging very quickly, and companies have to be right there to make sure they meet those growing demands.

Gardner: It sounds like you're really talking about risk mitigation, trying to find some focal point that allows you to support your legacy, move to the rich-client and SOA activities, as well as be ready to go to what some people call Web Oriented Architecture, and take advantage of these new hybrid deployment options. Does that sound like what you're doing?

Masri: That's a fair statement.

Gardner: James, is this something that we can expect to shake out soon, or are companies going to be dealing with heterogeneity -- not just in terms of technology, but in approaches -- for some time?

Governor: We actually see an acceleration in this area -- tools and apps that span clients and the Web. I’ve taken to calling it the "synchronized Web." How can you have two different sets of services talk to one another? In terms of how you develop in that environment, you’ve got to develop conversationally. It’s about message passing. Because of that, we all are going to see some changes around the language choices.

We're seeing some interest in terms of some interesting new development languages, such as Erlang and Haskell. We are certainly seeing interest from developers in those areas.

It's like enterprise software companies not having an open-source strategy. Basically, you need one. From an economic standpoint, you just don't have a choice. Any software company that doesn’t have a thorough-going strategy for understanding and developing both for Web modes and offline modes is really missing the point.

Whether we're thinking of our clients that come from Google Gears, whether we are thinking about offline clients using an environment like Adobe's Apollo Integrated Runtime (AIR), we're already thinking about spanning clients and websites.

From an enterprise standpoint, the same choices need to be made. User expectations now are that, they are going to be able to have some of those benefits and centralization, but they are also going to be able to have rich experiences that they're used to on desktop clients.

This is a very important transition and, whether it’s Pulse or any number of the Web apps we're seeing this from, we are definitely seeing this in enterprise Web development. It's really important for us to be thinking about the implications, in terms of the language support and in terms of runtime. We've already mentioned the Amazon Web services back end. We're going to be seeing more and more of that stuff.

There’s a little company called Coghead, and it’s really focused on these kinds of areas and it’s now excellent. They've chosen Amazon Web services as a back end and they've chosen Derby Flex as a front-end to give that interactivity. The Amazon model teaches, or should teach, a lot of software companies some important lessons. When I look at developers, certainly grassroots developers, it has almost become a badge of honor that you're getting, "This is what Amazon charged me this week."

The notion of the back end in the cloud is growing in importance again. That’s probably why IBM just announced yet another one of its, "Hey, we're going to take a billion dollars and move it towards cloud-computing" kind of initiatives.

Gardner: Right. We’ve obviously seen a lot of change in the market. Organizations and enterprises that depend on an ongoing evolution on a single-stack approach need to try to come up with the tooling and framework and environment that allow them to accomplish what they need from the backwards-compatibility perspective. They also need to put themselves into as low a risk position as possible for taking advantage of these dynamic environments and the change in the economics and the landscape.

We've been talking about the transition to WebSphere Application Server 6.1 and the implications for tooling, the pending arrival of MyEclipse Blue Edition from Genuitec, helping companies find some additional choices to manage these transitions.

Helping us weed through some of this -- and I have enjoyed the conversation -- we have been joined by Maher Masri, president of Genuitec. Any last words, Maher?

Masri: Just a reminder that the Blue Edition first milestone releases will be available in February. There will be a number of milestone releases that will be available for immediate access and we encourage people to download and try it.

Gardner: Very good. And, also James Governor, co-founder and industry analyst at RedMonk. What's your parting shot on this one, James?

Governor: Let’s get specific again. Some of this has been a little bit blue sky. I think it’s very interesting that IBM is has posted a pretty good set of financial results today.

Gardner: They're not going away, are they?

Governor: They are not going away. That’s exactly right. It used to be said that IBM is not the competition; it is the environment in which you compete. It seems to me that Genuitec and many others are probably a pretty good example of that. That was well put by you. IBM isn't going away.

Gardner: Well, thanks. This is Dana Gardner, principal analyst at Interarbor Solutions. You’ve been listening to a sponsored BriefingsDirect podcast. Thanks, and come again next time.

Listen to the podcast here. Sponsor: Genuitec.

Transcript of BriefingsDirect podcast on tool choices for WebSphere shops. Copyright Interbarbor Solutions, LLC, 2005-2008. All rights reserved.