OraTransplant

Oracle OpenWorld Continued…

25 September 2008 at 21:37 CEST | In Uncategorized |

There’s so much to see and attend at OpenWorld that it’s hard to make choices. I managed to attend a few other sessions. Out of personal and technical interest I attended a session on “Oracle Continues Query Language for Complex Event Processing”. This is something that’s not likely to be used in my day-to-day job, but I was interested to see all this new stuff.

Another session I attended was on creating a custom look and feel and skinning for the new ADF 11 Rich Client Components. This was a very very interesting presentation. Users aren’t always happy with the out-of-the-box look and feel of an ADF application. This was particularly true for 10.1.3 apps. Now with version 11 the default look and feel is so much better but it’s still good to learn how you can customize even that. In the end it wasn’t near as hard as I thought it was and as it definitely was in previous versions. One of the best things is that they finally documented all the skinning stuff thoroughly. There’s even a demo site with all components and some example customizations to their skinning. This gives you a quick and easy way into what it possible.

After that presentation I had my own presentation on OraFormsFaces, a toolkit to integrate Oracle Forms into a JSF/ADF application (or whatever other web technology for that matter). I won’t go into that for now as I’ll use a separate blog post for it.

Right after my own session I attended one from Oracle about the new ADF Desktop Integration feature. This allows you to build an “application” right in MS Excel 2007 with the same familiar JSF components one uses for building ADF web applications. You can just setup the “page” bindings manually in JDeveloper and then go into design mode in MS Excel 2007. It will show the bindings you’ve setup and allow you to insert a number of JSF components in your Excel sheet based on these bindings. Now you can have things like buttons and even fully functional ADF tables right there in your Excel sheet. When someone runs this they see live data from the data controls (which could be a database query, a web service, or any other ADF data control). A user can change data and submit it back to the server. Very much like a typical ADF web application but this time straight from the familiar Excel interface. For some (mainly bulk) data operations Excel could be the preferred choice for some users. It’s good to know a developer can offer this with the same JSF skills he already has for building ADF web applications.

In case you wonder; the Excel sheet can also be used without a live database connection. It will detect local changes and it’s up to the user when the data is synchronized back to the ADF server. This means it even allows for (simple) offline working with synchronization and conflict resolution later on. Lucas Jellema from AMIS told me that Duncan Mills previously stated that ADF will have full support for disconnected applications. Lucas blogged about that back in June. This would be so great as we have a heave Swing desktop application at Eurotransplant that is used by transplantation coordinators that are dispatched to hospitals. This application has to be disconnected in nature since mobile network access is not guaranteed at each site. Building and maintaining this application has proven to be quite complex and not far as productive as building ADF web applications. Hopefully this new ADF disconnected feature can make our lives much easier.

I finished the day with a presentation by Frank Nimphius on ADF security. I was expecting a sort of rehash of the same story Frank had for ADF 10g. However he explained that security architecture changed quite a bit with the integration of Oracle WebLogic in the technology stack. So he adopted his presentation to explain how things fir together with the new Oracle WebLogic. I must be honest that it wasn’t all 100% clear. To me it felt like it’s much more fragmented and requires stuff in even more metadata files. There was even stuff in some files that wasn’t actually used which adds to the confusion. There’s user definitions in multiple places and you’ll have to know which one is relevant. It all looked a little rough around the edges and it wouldn’t surprise me if this was one of the main reasons for not making the deadline of releasing JDeveloper 11g when OpenWorld started. They plan on having it out there next week, but I guess they did everything they could to have it there last Monday when OpenWorld opened.