View from the Edge

UK and Scandinavian PUG Events

We made a key part of our strategy for 2010 support for the OpenEdge community. We think we have a very compelling story with what we have delivered in OpenEdge 10.2B, and what we are working on in OpenEdge 11. Everything we are hearing supports that so we want to ensure that the story is getting out.

In that vein, we are providing strong support for the numerous PUG (Progress User Group) events that are being held around the world. In fact a couple of weeks ago I was in the London at the UK PUG event held in our corporate offices in Slough and in Lillehammer, Norway (O.K. It was actually Øyer) presenting to a combined Sweden/Norway Pug event.

Both events had 40-50 people, a nice collection of sessions covering a broad range of topics including several sessions by me. At both events I shared with the attendees the Progress vision for Operational Responsiveness, and introduced them to RPM (Responsive Process Management). I also reviewed with them the key features in OpenEdge 10.2B including Transparent Data Encryption, and introduced them to some of the key features in OpenEdge 11 including Multi-tenant Tables, Multi-tenant AppServer, and our thoughts around automated deployment to the cloud.

In Norway I also went into detail in terms of a number of best practices around building SaaS including multi-tenancy, UI flexibility, etc.

And there were lots of other great sessions as well. In the UK, British Airways gave a very comprehensive session on how they used Sonic, Actional, and DataXtend to build out a robust SOA infrastructure. Gary Clink showed how to achieve Business Transaction Assurance within your IT infrastructure using Actional and the AutoEdge sample application. There were other great sessions on SaaS, OO, and BI as well.

In Norway, Jarmo Nieminen gave an introduction to OpenEdge Architect, OpenEdge Replication, and OpenEdge Management.

Two of Progress' technology partners, Mike Fechner and Robert Prediger showed how OpenEdge is open and flexible enough that you can really use it with whatever UI technology you want. Mike showed how the OpenEdge GUI for .NET is great for building compelling desktop UIs and showed some tools that simplify migrating from the OpenEdge GUI to the GUI for .NET.

Robert showed how you can build AJAX style UIs for your OpenEdge application and showed off his Backbase-enabled framework to help you get there.

And there were other sessions on the latest Progress corporate positioning, and on application and database tuning as well.

But the PUGs are not just about learning and technical content. They are about socializing and networking with your peers in the OpenEdge community. It's about building the relationships that will help you to be more effective down the road at doing your job. In the UK they had a very comfortable dinner on the evening before the event. And in Norway they did something that truly fit in with the culture and environment that we were in and is certainly fits in with what I like to do best. I think these videos speak for themselves.

So if you are not a member of your local PUG, please join. And if you are a member, plan to go to the next event. Offer to do a session that you think will help the other members. Not only will you learn something, and help others learn, but you will meet a lot of good people, and have a lot of fun.

Ken Wilner

Ajax in Perspective

Don’t get me wrong. I am a big fan of Ajax. In fact I spoke on Ajax at the Progress Exchange conferences in 2006 and 2007, and was very excited to see that several customers were speaking at Exchange 2008 on how to use Ajax with OpenEdge. However, given all of the hype around Ajax, I was surprised to see a recent Forrester Research report, reported on by ComputerWorld, that Ajax is not exactly living up to expectations.

You see conventional wisdom would say that with Ajax, we have finally solved the rich vs. reach problem. No longer do we need to decide if we want our applications to give users a rich dynamic experience that is tightly integrated with the desktop like a traditional desktop application, or do we want to allow our users to have access to our application where ever they are without installing anything by building a Web (browser)-based application.

With Ajax I apparently get the best of both worlds. All I need to do is pick an Ajax library is write a little bit of Javascript code or better yet use a open source Javascript library like Dojo, Prototype, ExtJS, or script.aculo.us, to simplify the development process, and presto, I now have a rich user interface running in the browser just like it runs on a desktop.

And while that certainly would be great, according to Forrester, apparently things aren’t as simple as that. According to the Forrester report while Ajax is great for the “occasional user” or the “very infrequent user”, for the “power user”, that is the user that is using the application for a significant part of the day to do their job, Ajax just isn’t holding up.

 You can read the Forrester report, or the ComputerWorld article yourself to get all the details, but suffice it to say that the reasons why Ajax is not holding up is centered around rendering performance when building complex screens, and network performance because most Ajax tend to go to the server to do validation much more than they do in a desktop-based client-server application. And while many of the Ajax libraries do a pretty good job of hiding differences across the latest versions of the most popular browsers, the problem is not completely solved because many IT organizations use  several versions of many different types of browsers and incompatibilities still do exist even with the latest versions.

So, what does this all mean? Should you stay away from Ajax? Of course not. What it does mean, however, as Salvador Vinals points out in his post in the Progress SaaS blog, and as is discussed in the Forrester report, is that you need to pick the UI technology that meets the needs of the end-user, and if your application has multiple different types of end-users with different usage patterns, then it may be highly appropriate to support more than one type of UI.

So if you have occasional users or infrequent users, users that move around a lot, and need to be able to get to the application from where ever they are then you certainly should consider building a Web UI and using Ajax to get a richer experience. I certainly would encourage it.

However, if your users are power users using a complex UI and primarily using the application from one location, then using one of the more traditional desktop UI technologies may be more appropriate.

And because we know that needs of the users and UI technologies change fairly frequently, one of the most important things you can do is architect your application based on the principles and guidelines of the OpenEdge Reference Architecture so that you can more easily change your UI or support multiple UIs when the need arises.

 

Eclipsecon 2008

Last month, I had the pleasure of attending Eclipsecon 2008, Eclipse conference, in Santa Clara, CA. The conference was well attended with 1404 attendees representing 37 countries.  There were 10 parallel tracks and with a wide range of sessions from “Getting started” to “How to make money with Open Source”. In the past couple of years, the Eclipse ecosystem has grown rapidly with 11 top level projects that cover many of the areas of SDLC tooling. With the recent announcement of the Runtime (RT) project, they now have a significant traction on Rich UI client (using RCP technology), Rich Web UI (using RAP technology) and the application server areas with OSGi and “Server Side OSGi” projects. 

The conference in general had a lot of buzz about web as a platform for Eclipse and AJAX tooling. I have attended Eclipsecon for the past three years and the previous conferences have been very technical with lot of sessions on “how to” in Eclipse where they would get into code within the first 10 min of the session. But I noticed a slight change in focus this year – there were a lot of success stories, reference implementations, adoption related sessions and many business focused sessions like “How to make money from open source”. In my opinion this change somewhat reflects the maturity process for the Eclipse ecosystem.

You can browse thru’ the conference materials at http://www.eclipsecon.org/2008. Happy reading!

So in summary, Eclipse has gained significant momentum and looking at the future direction of where Eclipse tooling is headed further strengthens OpenEdge Architect strategy of providing an open and productive tooling platform for OpenEdge customer base. The continued growth of Eclipse ecosystem will continue to enable OpenEdge customers to take advantage of the Eclipse platform and ecosystem.

One of the goals for OpenEdge Architect is to make OpenEdge n-tier business application development iterative and more productive. Looking at the different Eclipse projects, there seem to be definite synergies between Eclipse projects like Web Tools Project (WTP) for providing AppServer and Webspeed tooling support, Mylyn and the need for task based development environment, Dynamic language toolkit (DLTK) makes it relatively easy to add more editors feature quickly, Data tools project (DTP) for database schema navigation etc…. So we are looking to further prototype and explore of some of these Eclipse projects to meet OpenEdge customer needs.

Following is a brief summary of some of the sessions that I thought were interesting - more details about this can be found in the conference material on the link above.

1. Eclipse 4.0 (e4) – There was a lot of discussion about the next generation of Eclipse tools – Eclipse 4.0 (e4). E4 is expected to ship in June 2010 with a prototype ready by June 2009. Web as another platform for Eclipse and SWT was one of the main themes for Eclipse 4.0 platform team. The platform team had a few cool prototypes and demo on possibilities. Eclipse committee seems to be still undecided on the definite direction on how Eclipse will be available on Web. Some of the AJAX solutions that were being considered included RAP, GWT, DOJO library with Java compiled to JavaScript, Flex and Action Script, Silver Light etc…

2. Rich AJAXPlatform (RAP) – RAP technology is very promising and provides Rich UI on Web for applications that are built for desktop. The RAP platform takes an application that is built using RCP (Rich Client platform) and renders it on the Web – without many changes to the application source code – single source multiple platforms. RAP dynamically exchanges SWT widgets with RWT widgets that work in client-server distributed mode.  Branding for the RAP based web UI provided by Themes.

3. Eclipse Dynamic Languages Toolkit (DLTK) – Very promising technology for adding new editor features to IDE’s. The DLTK relies on a standard “Structure Model” to provide functionality like outline view, search, open type, open declaration, documentation access, code completion. Internally DLTK has a selection engine that provides a lot of capabilities mentioned above.

4. Mylyn – The task based development paradigm promoted by Mylyn has lot of traction and I heard lot of developers starting to use it. The Eclipse 3.4 release seems to further add more stability and extensibility and a host of other SCM (software configuration management) tools are integrating their solutions with Mylyn.

5. Data Tools Project (DTP) – DTP functionality keeps getting better and better – they provide Data source configuration, navigation, SQL  editor, SQL query builder etc... DTP’s real value is in its integration with other projects in the Eclipse ecosystem – like BIRT which makes the story very compelling.

6. Security in Eclipse 3.4 – The security improvements in Eclipse 3.4 are significant. Trusted bundles, ability to encrypt sensitive data in Eclipse workspace and Certificate store in RCP world are some of the enhancements to the Eclipse 3.4 release.

7. JMaki – This open source client-server AJAX framework is mainly driven by Sun and the Glassfish guys and conforms to Open AJAX standards. I was very impressed with the demo and it appeared to be very easy to build Mash-ups. This has Eclipse integration. This framework makes it easy to use most other AJAX toolkits (DOJO, etc…) and it acts as a standardized layer below the AJAX toolkits. More details at http://jmaki.com