Web Dynpro for Open Source

Question

Dear Web Dynpro community,
please have a look to Benny´s  blog under http://www.sdn.sap.com/irj/scn/weblogs;jsessionid=(J2EE3417200ID0731044050DB00226095944409325288End?blog=/pub/wlg/16004 raising the question whether Web Dynpro Java should become Open Source.
It would be very interesting for me to here your opinion about this. What are the benefits in doing this? What are the risks? Would YOU like to actively contribute?
Thanks in advance
Jochen (Product Owner Web Dynpro Java)

Answer 1

Hi all,
<p>
let me add some comments on some parts of the answers (also in the two blogs around this topic ..).
In general it is really interesting to see that there seems to be a real interest for an "open sources" Web Dynpro for Java. This is motivating to go into this direction and to further discuss this here within SAP.
<p>
Some technical comments: I agree that an open-source WDJ (in the ideal world) would include also the WDJ tools parts and most probably also UR ( = Unified rendering) as the underlying (HTML) rendering technology. Also NWDI is somehow part of this game. BUT: I do not think that such a big step is realistic and feasible. Instead of this the first natural step for me would be to open-source the WDJ runtime part. This WDJ runtime contains all important parts needed to run WDJ applications (including the control libraries, the different CMI model implementations, the different clients and last but not least the server-abstraction layer).
<p>
This server abstraction layer (SAL) is most probably not know by the WDJ community as this was never rolled out in the past. As the name shows this layer encapsulates the server environment "hosting" the WDJ runtime. This allows us even today to run WDJ applications besides the SAP JEE  also in top of a simple TomCat installation. This version is used for a long time for control development but was never rolled out.
<p>
In the labs we have WDJ application running on Amazon Cloud, on Google App Engine, on Bea and also on top of Geronimo.
Not all of these server abstraction implementations are fully working for the moment and not all features are available for all implementations - but it would be really straightforward to provide all these implementations - in the ideal case of course via a active open source community ,-)
<p>
Besides the platform implementations I see some other areas how an open sourced version of WDJ could be extended:
<p>
The combination of WDJ projects with Maven. This is also on our internal development list but due to prios this was never started.
New controls. Also without open-sourcing the UR layer I see a lot of potential to extend existing controls or to build completly new controls.
"Generation-free WDJ" - the idea would be to provide a WDJ version which is not depending on a complex generation framework. This would be of course one option but a "generation-free" WDJ would allow you to code your WDJ apps / components completly without any generation.
Extending the usage scenarios for "system components" - this kind of WD components is used to extend WDJ runtime capabilities. Currently this is only used for the "Technical Info" and the UI of the implicit personalization but there is a lot of potential to extend this.
<p>
This is only a very short list of things I have currently in mind - for sure you will find more.
<p>
Have fun at TechEd Phoenix
Best regards
Jochen