It has been a while. But now we are finally about to witness the production release of JDeveloper 11g and ADF 11g including the ADF RichFaces components. Today, Oracle announced that JDeveloper 11g will be available for download from OTN from October 1st, in a little over a week. At that time, ADF 11g application can be deployed on production systems and will be supported. Note that ADF 11g at this point is certified against WebLogic 10.3 and currently does not support OC4J. JDeveloper 11g ships with WebLogic 10.3 as its integrated web server; the embedded OC4J container is no longer there.
The JDeveloper 11g release we will get access to on October 1st is duped ‘the Boxer release’. Sometime in the first half of 2009 we will get the successor, called ‘Bulldog’. The Boxer release does not contain the SOA Suite design time – as the SOA Suite runtime will not be available until some time next year – nor will it contain WebCenter. There have been suggestions that it does not contain the MDS and Active Data Source features either, nor the Hierarchy Viewer..
Currently the ADF Runtime license is free for organization deploying on WebLogic (and OC4J 11g when it becomes available). For other application servers, the license is about $120/user, $5800 /processor. The ADF runtime license now does not just include ADF 11g Runtime but TopLink 11g and MapViewer (needed for the geomap DVT component).
The Bulldog Release – the road ahead
First of all: JDeveloper 11g was originally slated to synchronize with all of Fusion Middleware 11g. However, partly due to the BEA acquisition, the release plan for FMW 11g has been revised. The current word on when FMW 11g will be released is first half 2009. Since the Bulldog release for JDeveloper is mainly about synchronizing with FMW 11g, its release date will coincide with that of FMW 11g.
Bulldog will first of all provide the design time for all of FMW 11g, including WebCenter and SOA Suite.
Other areas that we will see new or improved functionality in:
- Application Life Cycle Management – support for collaboration in development projects including Interfaces to Bug Tracking, Project Managementt tools (MS project), Requirement Management and interface to Integration Server (Oracle Team Server) [a new term for me]. (I was wondering whether this also includes a link with Oracle Application Test Suite?)
- Mobile: inlcuding disconnected clients (not based on Google Gears); Oracle acquired the assets of AppForge some time ago, a company that did a lot on data synchronization for disconnected clients that upon reconnection needs to be realigned with the server; their product – Replay Server – is likely to play a central part in this
- ADFdi – desktop integration (licensed as part of ADF, no additional cost for Oracle AS/WebLogic customers); using Excel as another ADF User Interface, including off line functionality
- ESS Base (Hyperion) Data Control
Open Source – Trinidad and ADF
Oracle had previously announced that ADF Rich Faces would be donated into the Trinidad Open Source project. That is going to happen in the very near future. In addition, there are very serious plans to also donate the ADF Faces Controller to Trinidad. That would mean that Task Flows for example become available to the open source community.
The JSF standards are not very well defined in the area of Controller. There is the Lifecycle obviously and the Navigation mechanism – which is fairly bare bones. There is little support for reusing/nesting fragments in an intelligent way, having these fragments walk through their own navigational logic etc. By donating the ADF Controller to the open source community, Oraxcle hopes to help shape the near future of JSF (the 2.0 release of spec perhaps?) and make sure that the controller side of JSF is strengthened.
Browser Support
ADF Faces is certified against the following browsers:
- Firefox 2 & 3
- Internet Exploirer 7 (IE 8 testing right now)
- Safari 3.0
There is no support for IE 6 – nor will their be. Testing Google Chrome is being conducted at this moment; there are some (minor) issues with the JavaScript V8 engine in Chrome.
In terms of performance, ADF Faces has the best (client side) performance on Safari. Firefox 3.0 is substantially faster than 2.0 – about 25%. Internet Explorer 7 has similar performance as Firefox 2.0. No metrics are available for IE 8 at this stage.
New initiatives
In the very near future, the Oracle ADF team will make a live demo available of the ADF 11g Rich Faces components. The url for this demo application is: rea.oracle.com .
Another initiative is currently duped ‘CodeShare’. It is to be a site where samples, skins, utilities and other useful reusables created by the ADF community can be collected. It coulld be something like a SourceForge project or an area in Google Code. What shape it will take is currently being decided.
Speaking about the ADF Community: over 40 ADF specialists from around the world have joined forces in ADF Community (see The world wide ADF community – convenes in the Oracle Wiki, develops the ADF methodology and meets at Oracle Open World for more details and the urls for the Wiki and the Google Group). The community convenes during OOW at the Unconference process to discuss ADF obviously and brainstorm on how to make the community work in the near future. Oracle will be present – representatives from Product Management and – at least as interesting – developers from Oracle Fusion Applications development who have been working with ADF 11g for the last year or so.
Interesting discussion.
If I can add my word as a non ADF expert, i would say it’s clear we have two approaches here when talking ADF :
one from a Java developer expert developing for customers and with sometimes a technical choice driven by the customer choice and the second is an approach from another Java developer expert but speaking more from a customer perspective.
As a company CIO, this difference means a lot to me specially if you want to own,master and sustain what’s running in your company. That’s why I understand the choice of Florin’s customer for JDev 11g.
Apart from rich functionality and RAD, what Oracle has to deliver with ADF is sustainability of apps built with the framework by insuring ease of migration from version to version.
ADF 11g will be a great release, the mentioned CodeShare initiatives combines with an “ADF Methodology” would be great stuff for ADF developers.
Thank you Lucas!
Hi Peter,. On your question with regard to SQL Developer: I asked Sue Harper, product manager for SQL Developer what would be shipping in JDeveloper 11g and she confirmed that JDeveloper 11g production will ship with SQL Developer 1.5.2 (one patch release later than the current shipping release of SQL Developer itself.
I am located in Romania, unfortunately not close enough to your location (Holland if i am not mistaken).
Florin
Sounds great. And I am sure interested in your real life experience with ADF 11g. Where in the world are you located – close enough for us to meet in person?
Lucas, thanks for your answer, I really appreciate it.
We switched from Open Source technologies to Oracle because that was one of our customer decision. The choice of using JDeveloper 11g was a customer decision also. JDeveloper 11 and ADF seems a very bold goal, therefore risky. I hope that Oracle manages to deliver something meaningful outside Oracle world. Lets resume this conversation in December, after 2 months of working with production ready release.
Florin
Peter, I do not exactly know the status of SQL Developer for the 11g Production release. I would expect a slightly older than the very latest SQL Developer to be in there, but I cannot be sure. I will try to ask one of the JDeveloper product management team here at OOW. And perhaps they may read this comment and come in to answer it themselves. I will get back to you that one.
Hi Florin, Thanks for your comment. It is very interesting how you state all the frameworks you mention have made you much more productive than the ADF stack – and yet you have switched to the ADF stack. That does not seem a very good idea then. Or can it be that you have a different type of job to do now and you cannot really compare – the apples and oranges thing? You are right that in theory you should not have to be aware of the XML because of the IDE and the wizards and you are right too that sometimes it is unavoidable to dabble in XML like for example the faces-config.xml file for JSF. One of the benefits we get from the XML – apart from having the wizards and editors and a lot of other IDE support like impact analysis that most annotation driven environments do not seem to give you – is the ability to perform run-time manipulation of the meta-data either before or after it gets loaded in. That for example makes it possible to apply organization or user customizations in the case of SaaS style applications.
One of the undercurrents running through JavaOne 2008 by the way was the notion of the ‘annotation hell’ – using one or two annotations in a class definition is much easier than having to set up yet another external XML configuration file, especially if you do not have an IDE that helps you edit the file or shields you from making typos. However, adding dozens or even many more of annotations to your classes quickly render you in their own little version of hell – readability, maintainability, ability to perform bulk changes on those settings or get some across the board reporting are all pretty much out of the window.
I am afraid I apparently was not clear enough in my previous writing: I did not intend for you to understand that the developers that are actually developing JDeveloper – the product development team itself – has people with a lack of Java skills (duh) but I tried to describe how JDeveloper is being used by 1000s of developers in Oracle’s other product development teams – especially the horizontal and industry specific ERP applications – who may not all have a strong background in Java (or even in programming for that matter). One of the challenges for the team that creates JDeveloper is deliver a tool that on the one hand allows people like yourself to do all standard Java and JEE programming, either using fancy IDE support or just doing your thing in source code, and at the same time help those with fewer skills and experience to be productive too. And you will find many teams where developers with a limited skill set are turning in their fair share of work, with the more advanced stuff beyond the IDE and the wizards capabilities being picked up by more senior developers, like yourself.
I think by the way that it is quite brave to pick up such an early Technology Preview to start building an entirely new application. There is a world of difference between TP3 and the pending production release – and obviously what you have done is certainly not what Oracle (officially) wanted you to do with the early pre-beta. Apart from the pre-production issues that I am sure you have run into, are you satisfied with the application that the team is producing? Is ADF giving you what expected it to and what made you decide to start using it back in February despite the fact that it was not even beta at the time?
Lucas
Lucas, I am fully working from February with JDeveloper 11g. I have 6 years of java open source background (JBoss Seam, Struts, Struts2, Tapestry, Hibernate, Spring ). All open source technologies or frameworks I have mentioned here made me and my team much more productive than the ADF stack. One mistake is this huge XML configuration that can prove overkill . When others are running from ‘XML HELL’ (in favor of annotations), Oracle embraced it. In theory, XML configuration won’t be a problem since you always use the IDE and wizards. But in practice it won’t be always the case, you have to understand what is happening under the hood when you are working with production applications.
You mentioned those not experienced with java developers that work on JDeveloper. Now you confirmed me what I was suspected for a long time: some of the developers on ADF and JDeveloper don’t have the minimum Java knowledge required. Example: handing Exceptions. Anyone who worked with JDeveloper 11g for some time knows what I am talking about.
I hope that the Production Release will prove that all my bad experience was because of the Technical Release status of the JDeveloper.
Florin
Hi Lucas,
Would you know if SQL Developer will be available as a plug-in in the 1 October JDeveloper release?
Regards Peter
Andrej, thanks for your reaction. Could you delve into what these better solutions are – for which they are solutions and perhaps the criteria on which and the circumstances in which they are better? You are making a fairly broad statement here – which Oracle (BEA) products are you referring to?
Personally i think oracle is going the SAP way. Nobody uses SAP netweaver for new product development, only to modify existing SAP applications, eventhough Netweaver is J2EE complient. I think the future oracle stack developers will be modifying oracle applications, but the stack will not be used by companies that do not use Oracle Applications. Not good enough, too complex, too expensive. There are better solutions out there if you’re not tied to Oracle Applications.
Guys, you sound more and more like Statler and Waldorf (see Muppet Show). You may not like (all of) it – and obviously you do not. But your comments are pretty negative.
Of course Oracle is not moving as lightningly fast as certain smaller organizations and open source communities are. And yes, the acquisition and current integration of BEA products has slowed things down. And the Oracle product development teams have been focusing very much internally on the 1000s of developers developing Fusion Applications and other Oracle Applications products – sacrificing time to market for robustness, productivity and enterprise level quality. Among the 1000s of developers in Oracle development teams many are not as proficient in Java and JavaScript and other web technologies as you may like them to be. They come from very different backgrounds (Oracle Forms, other development technologies or even no software development to speak of) and still need to be productive. ADF is very much focused on delivering 90% of the functionality required in most applications in a very declarative, productive, virtually code free way- allowing the vast majority of developers who are not top notch Java developers to pull their weight. The next 9% will take some considerable programming skills, experience with various Java & Web technologies and quite a bit of creativity. It will not be as productive as the first 90% obviously, and you may not have the flexibility you two would like to have. Then there is 1% desired functionality that is virtually impossible or at least extremely expensive to develop.
I am not sure what you mean by too little – too late. I would have preferred to have some components a lot sooner than I am now getting them. At the same time, I am happy with the enormous impact the experience of the Oracle Application development groups has had on the way the technology has evolved. ADF and JDeveloper are all the more robust, usable and productive for it. What is ‘too little’? What do you want/need that Oracle is not giving you and you cannot use from other vendors and integrate with Oracle’s stack? You may not be totally up to speed with everything on offer so perhaps the situation is not as bleak as you perceive it to be.
Looking for innovation – now there is an interesting discussion. I get often accused of always wanting something new. The latest technology, the current trends both technology wise and in terms of architecture, end user experience etc. And I am using Oracle technology most of the time. What kind of innovation are we speaking about? My main goal is providing my customers/end users with the functionality they require in a context of non-functional decisions and choices that satisfies their longer term plans. Innovation is not typically I want to push. I want to explain the options, discuss the pros and cons and arrive at a balanced decision based on the actual requirement and context. Not drive the discussion based on my personal desires for applying the latest and greatest technology.
One thing I have come to realize over the years is that not all developers are created equal. What I find hugely challenging yet very satisfying can be frightening the hack out of someone else. The majority of developers is not me (or you for that matter).
Too little – too late.
Sounds very much like oracle, soon we will have something very exciting, just wait another year, and by then will postponed it some more… But if you need innovation now, better look elsewhere…
Here we go, Let’s build great apps.
And next year the next release with great new options like disconnected clients and ADFdi