Our talk in entitled: Exceeding User Expectation with JBoss Enterprise Portal Platform

Who are users? Employees, customers on smartphones, business partners, those calling your service desk? All of the above? Way back when, it was OK to have your internal users using a portal, your external users using a content management system (CMS), and your customer service reps using a web app. However today, siloing your user experience is inefficient and overly complex.

Join Red Hat’s middleware team in charge of delivering next-generation user experience products to learn how current and future Red Hat technologies can help your users’ experience. And learn about how other Red Hat customers are changing their end user experiences.

I will explain how EPP6 will be a step forward building websites more easily and in particular talk about how we’ll handle mobile support, I will also have the opportunity to announce a new opensource community project around GateIn Portal so stay tuned and expect activity on gatein.org.

Our talk will be followed by a Birds of a Feather open discussion, I hope to meet many of you there for a more informal discussion about your expectations around EPP.

I also won’t want to miss Thomas Tik’s presentation about OpenSource usage of EPP within Allianz, I’ve heard of various project there using EPP but I still have more to learn, the talk in on Thursday 28th at 4:50pm

In this session, Thomas Tik from Allianz, the world’s largest insurance company, will lead you through the creation process of an open source software strategy. Thomas will discuss the current state and future direction of Allianz’s open source strategy, sharing many real-world examples of the architecture definitions that resemble that strategy.

Thomas will also discuss the company’s use of JBoss Enterprise SOA Platform, JBoss Enterprise Portal Platform, and Red Hat Enterprise Linux. And he will end the session with a discussion of the success factors for large-scale, open source adoption.

The technical scope of this session ranges from app- and portal-servers, WEB, rich client, and SOA frameworks to operations and rehosting.

Flying on the Sunday, I will be on time for JUDCon which will happen Monday 25th and Tuesday 26th and be happy to meet people, other members of the GateIn/EPP team will also be around at JUDCon: Nick Scavelli, Ken Finnigan (our PortletBridge Guru) and Marko Strukelj.

The release note is available here. But before I go into 5.2.1 let me step back a little bit as I didn’t communicate on the release of EPP 5.2.0.

In EPP 5.2.0 we did major improvements the list of the new features is here but let me highlight a few:

• The URL handler is now customizable, it allows to have more customizable URLs, it also allows to have parameters in the URL, we use it internally for EPP SP to include the language in the URL such as http://www.example.com/fr/something. Additionally we got rid of the /private and /public part in the URL (which can be re-enabled for backward compatibility).
• A performance bottleneck with handling a large number of navigation nodes has been removed, additionally first login time has been drastically decreased (could be an issue in certain scenarios)
• We introduced a way to declare a portlet filter to be applied by default on all portlets, this was particularly interesting to add analytics or logging filters to portlets. We use it internally for an EPP 5.2.1 feature that I will describe just after.
• WSRP over WS-Security is now possible
• Ability to use managed or non-managed transactions (and improved in 5.2.1)
• Portlet surroundings can be disabled by default, it makes it easier for users who don’t want to have the “boxed” look on their website
• Dashboards objects are now created only if dashboard feature is used, this drastically reduce the database usage for customers who don’t use the dashboard feature or when only a small amount of overall users is using it
• The HTML markup has been simplified to some extend. We tried to keep existing skins working out of the box after a migration but you may need to adapt it. This change was required to save about 20% in performance when a markup is sent on a network (saved on markup size) and when a browser interpret it (saved on complexity).
• XML descriptors has always been an option to describe a site and its pages, after the first start all changes had to be made through the UI. In EPP 5.2 it is possible to always use XML files and modify pages without using the UI (require a restart). See the reference guide for more information.
• Ability to export and import sites through admin UI, REST services and command line
• And much more…

5.2.1 also received a set of bug fixing, performance but also self analysis internal features to recover from failures, a faster JCR reindexing…

We hope you’ll enjoy this release and we recommend everyone to upgrade to this release.

As usual EPP 5.2.1 and EPP SP 5.2.1 are available on the Red Hat Customer Portal.

Site Publisher enhance the core portal with Web Content Management features, it allows to write content directly from a website, get it validated, manage content such as web pages and images…

The full release notes can be found here, but here is a small overview of enhancements:

• Last-modified HTTP header correctly set when serving resources
• Configurable security domain
• Various fixes for Internet Explorer 6, 7 and 8
• Closing a document that is not saved now raises a warning message
• Content portlets were duplcated after clicking the Import Applications button in the Application Registry page
• And tons of bug fixes

It is recommended to all users of EPP 5.0.0, 5.0.1 and 5.1.0 to upgrade to this latest release as it contains an important security fix along with bug fixes and performance improvements. Also for support reasons you will be asked to be running the latest available version. This release is backward compatible with previous releases of the 5 series.

The release notes are available on docs.redhat.com, but here is a short summary for some of the issues:

• A security issue filed as “Important” has been found in the Seam libraries shipped with earlier versions of EPP, you can see the details in the Red Hat Security Errata, in the Red Hat CVE database in the MITRE CVS database or on the NIST National Vulnerability Database.
• Few potential XSS issues have been fixed (mostly in the administrator part)
• CAS token validation will now work over HTTPS
• Fixed issue in SSO integration with charset encoding
• From now setting up a context root to /something/somethingelse will now work, in previous releases, having a “/” in the context root would have failed.
• When using ExpirationAlgorithm for Picketlink cache configuration, the entries from the cache would never get evicted, this has been fixed.
• WSRP consumer configuration file specified in wsrp-configuration.xml is now taken into account.
• In the calendar popup, the translation wasn’t possible, it is now possible and available in French
• Major improvements in JCR index storage consistency and performance
• Javascript compressor can be turned off without throwing ERRORs.
• Facilitated the configuration of multiple clusters on the same network.
• …

This project is key into the strategy of the JBoss Enteprise Portal Platform as it makes it easy for any developer to use JSF, Seam and Richfaces applications inside a portal. Later, we will also want to support more web frameworks (to be defined).

Next step for the project is to fully support Richfaces 4 and JSF 2 as part of PortletBridge 3 and in the future find new ways to simplify the development of new portlets using different web technologies.

This job requires high software engineering skills to maintain the existing and develop new features, a strong motivation in supporting our customers and an in-depth knowledge of JSF internals and Portlets.

The official job announcement can be found here and is open in any country were Red Hat as a legal entity.

Come join Red Hat !

Portals are dead.

My oh My !

Once that myth busted (but portals have a long history of having to justify themselves so I won’t be done anytime soon), I moved on to a myth I heard from some people that a portal == MyMSN, MyYahoo, MyAOL, iGoogle… as a matter of fact, most portal vendors let you give to users the option to personalize their experience, it’s really just an option that doesn’t always make sense (we’ve seen project failing as they spent an incredible of time working on dashboard, when customers would not want to spend any time custmizing their visit (like when visiting a an insurance company to get a quote…) In fact most of our users don’t use the dashboard/boxed approach -> Myth Busted

Too complex/rigid ?

“Portals are too complex and rigid” say some people, this one is interesting as like any other framework, there is a time to invest to learn how to use it, and the hope is that the time saved later on will compensate it. All servlet tips’n'trinks won’t work with a portal, but with technology bridges like the JBoss Portlet Bridge *some* of those differences from traditional Servlet to Portal environment are abstracted and hidden from the developer which helps. When it comes to rigidity, we (as portal vendor) take care about being able to integrate with existing infrastructure, we won’t expect the infrastructure to change. -> Partial (Learning curve)

I’d rather create my own !

Other myth is that it’s better to create its own, I won’t spend much time on it, there is always this religion war between people who feel safer when they build something vs people who prefer to reuse proven frameworks. Their is also some flexibility as most people will not think of writing down to assembly because they don’t trust the language, to the other extreme some people will be satisifed with out of the shelves solutions.

Portal is a must have

The last myth is actually a joke, as I do not think any framework can save all problems obviously, I just wanted to have a “Confirmed” and at the same time, crack a little joke and remember my not-so-unbiaised situation.

JUDCon May 2nd and 3rd 2011: A developer oriented conference at a very affordable price (99$). You will get to hear about latest and coming development technologies, here are few sessions I picked which should be of interest for the audience of this blog:

• Develop Rich Web Applications with Seam, RESTEasy and ExtJS – While not targeted to the portal environment, the applications developed with such techniques could be reused in a portal environment thanks to the JBoss Portlet Bridge.
• BoxGrinder – Your Appliances in the Cloud – We already built images of our project (GateIn) to run on the cloud thanks to BoxGrinder. This could be the future for EPP !
• Hacking Java EE: CDI extension n00b to l33t in 60 minutes – Let us know what you think of CDI ! We are looking into CDI at the customer facing technology to access our services.
• The Mobile Web Revealed For The Java Developer – Because it’s coming !
• Filling The Gap : Going Mobile With JBoss Technologies Today – Co-Presented by our own Wesley Hales, GateIn/EPP developer and JBoss Portlet Bridge lead !

So don’t forget to register for JUDCon, and if you want to meet Wesley, let him know.

Following JUDCon, we’ll have JBoss World, early bird pricing ends today ! It’s also in Boston to make it convenient to join both events.

JBoss World has 53 sessions plus 71 for the linux talks but the two to not miss if you read this post are:

• “Making Portals cool: The compelling advantages of a portlet bridge” which will be again presented by Wesley, our JBoss Portlet Bridge lead, here is the abstract:
  

  Portlet Bridges (JSR-301/329) have created a firestorm of welcomed publicity to portal technologies. In this session, Wesley Hales will show you how to leverage your existing investments in JavaServer Faces (JSF) and supporting technologies and put them to use within a portal environment.
  Most JSF developers do not realize that their JSF application can be easily deployed to a portal with only two jars and an XML file. In this session, you will  learn more about the possibilities of a portal combined with JSF, JBoss Seam Framework, and RichFaces, as well as why you would even want to use one.

  Attendees should have a basic understand of JSF and will leave with a new skill that can be added to their resumes.

  Wesley also introduced his talk in his own blog.

• “JBoss Enteprise Portal Platform BoF” co-presented by Jason Andersen and myself. Here is the abstract:
  

  Over the past two years, JBoss has introduced an entirely new JBoss Enterprise Portal Platform to the market, marrying the best of open source portal technologies with new capabilities such as embedded content and site management. After providing a quick behind-the-scenes view of the past two years, Jason Andersen and Thomas Heute will provide a glimpse into the future of JBoss Enterprise Portal Platform. For example: Will it still be a “portal”? How will JBoss and GateIn incorporate new computing platforms (clouds) and technologies (HTML5)? Join product management and engineering for this interactive session to discuss answers to those questions and more.

Again, the early-bird registration ends today, so be quick if you want to save some money on the registration !

JBoss.com hosts a page that explains the main differences for all JBoss Products, it includes a video, a diagram taking the Enterprise Application Platform as example and a table as shown below:

What does it mean for EPP ?

OpenSource

The project and product are both fully available under the LGPL License and sources are available to everyone.

Releases

There is always a cursor to set between “bleeding edge” and “rock solid” the following summarize the difference from a project and a product from a release point of view.

Like many other open source projects, GateIn Portal is always in a fast pace development mode. Developers of the project are in charge of the releases and always make sure all testsuites passes prior to a release, with all the passion brought by developers testing is never enough and automatic testing has its limitations.

On the other end, an EPP release has a stabilization period of several months. Soak testing, clustering testing, performance testing, UI testing… takes time and during the stabilization period, we only add bug fixes, no new feature, all that work of testing is done by a dedicated team  with automated and manual testings. Obviously in that timeframe the product also benefit from the community feedback as the features are already available in the project.  The community participates in validating the quality of the product, this is a huge help, something that commercial product don’t have (except with Beta programs) !

Certifications

As part of the testing phase, we are able to certify a list of thirdparty applications/servers, so you don’t need to worry if you will be using Microsoft Active Directory and/or CAS and/or Oracle Database… See the list here.

Support

The other very important aspect of EPP (the product) is that you get it with a support contract and a SLA (up to 2 hours response time 24/7) ! No need to post in the forum in the hope to get an answer, with the support contract, Red Hat has to answer your question and we take support very seriously, it is what we sell and we cannot offer to be bad. As a matter of fact JBoss support has always been ranked really high and Red Hat was in the top 2 vendor value for the past 7 years !

A particular release of EPP is supported for 7 years with patch release and one-off patches for production issues that can’t wait.

And more…

On top of that the EPP product comes with a legal protection, you don’t need to worry if one of our developer or contributor included code incompatible with the LGPL in the product, the customer is covered. The product comes with a configuration called ‘production’ which is a better start for a production system (in terms of configuration).

I hope to convince the EPP Product Manager to join me to share his thoughts about the portal landscape and our offers. We’ll also announce and detail our partnerships and talk about our WCM solution with eXo.