I’ve posted what is now very nearly an exact copy of my Extreme Reusability series (Part I, Part II) to the ADF Methodology section of the Oracle Wiki here. Why am I putting two separate copies up on the web?
Well, I’m going to use the copy on my blog as a version of Extreme Reusability that I have control over. I’m sure my understanding of the methodology will evolve over time, and I’ll continue to post updates as it does; I also hope that you all will continue to post comments on it here. The version on the Wiki, however, is, more or less, a donated snapshot to the community. All I particularly ask is that people maintain the links I’ve put in it to the two blog posts where I first laid the methodology out. Other than that, the content is fair game for anyone who has ideas, responses, etc.
I’m sick, so no actual substantive post this week–this’ll be it until next Monday.
Brad Worsfold | 29-May-09 at 9:19 am | Permalink
Hi Avrom. I was wondering if your extreme reusability can be done in 10g of JDeveloper? I am not able to see EO or VO objects from JAR files. The only way it seems to work is via source file inclusion. Am I missing something? I like the concepts, implementation is a bit of a challenge.
Avrom | 29-May-09 at 2:22 pm | Permalink
Hi, Brad. Your answer is, “well, sorta a little bit.”
You can certainly use the “Generalize, Push Up, and Customize” technique that is part of ER–creating framework classes, custom components when appropriate (although IIRC you can’t create custom components declaratively in 10g), etc., and deploy these as ordinary (Java class file-only) JAR files that can be used as libraries. And you can import business components, not as libraries, but as uncompressed (but compiled) packages using the File | Import functionality.
But 10g has no concept of reusable applications, and honestly the import functionality is a little bit wonky. While I encourage you to Generalize/Push Up/Customize whenever possible, Extreme Reusability as a full development methodology is going to be pretty hard to put into effect without the 11g technologies that, honestly, inspired it: BC libraries, application libraries, task flow calls, declarative custom components, templates, and so on.
I wish it were otherwise, because I’m not going to glibly tell you to upgrade–I know that for a variety of reasons (investment in Portal, etc), upgrading may just not be feasible for you. But I think you have to regard the usability of this methodology as another piece of 11g bling.