JDeveloper, ADF, lack of support and bleak future
552183Dec 17 2006 — edited Dec 18 2006Hi,
I'm a seasoned Java and .NET developer, used to many environments and frameworks. A happy user of Oracle databases.
For a very Oracle oriented client I started with a team of developer a project using JDeveloper (no choice). After 4 months of using JDeveloper, ADF to build web sites, here are some thoughts I would like to share. This remarks only focus on that products, because I have a lot of respect for the Oracle databases and had a lot of success with them.
I am starting to get depressed about the very few blogs (most from Oracle) talking about JDeveloper and ADF, the lack of books, and the Oracle forums.
Moreover, most of the time you never get an answer to your problems, sometimes untested not working 'solutions', and a very very few times a working solution. Help is sparse, the community is missing.
If Oracle wants to promote its framework, it has to boost the number of tutorials, examples and support.
Googling about ADF is most of the time a non sense since you always land on the same 3 blogs and the Oracle forums.
Proof of non use by the community ?
If you try on www.indeed.com to find statistics about JDeveloper or ADF jobs you get another frightening proof.
Sometimes you find a piece of code or a tutorial that may help, but you realize it works only on some specific release of JDeveloper.
What makes me angry is when you ask why your code works on the x.x.x.x version but not the newest version, and the only answer you get is stay with x.x.x.x version (I won't bother to fix your 10 lines or code or even give you a direction to do it (maybe coz I don't understand myself were the breaking changes are)).
JDeveloper is damn too slow ? You should have a dual-core and at least 2 Go of memory, you poor thing ! Is Java for rich people ?
Debuging is a nightmare. Compared to Eclipse or others it is not practical at all.
I'm happy when I can create a few pages drag and dropping controls. Business components are also a good thing. It speeds up my development. But since development is not just UI and generating business components, I lose a lot of time every time I want to write code, every time I do something "exotic" it becomes a pain in the ass.
My problem here is that the architecture and JDeveloper constraint the developer a lot. I don't even talk about the fact that everything is only Oracle database oriented (openess ?). Even changing your page layout is a pain. Positioning controls is non trivial. As a whole, I would say that this whole thing imposes you to adapt to it, it doesn't adapt to you and your way of working.
Most results I saw look like the Oracle websites.
I could go on about the various bugs and problems that plague JDeveloper. Like the refactoring (when usable) that leave old and annoying stuff in your model XML, like the bugs in the diagrams, like problems related to auto-binding on the page and manually changing controls ID, various synchronization problems between pages and backing beans, adding upload is a mess, OrdDomain types are buggy, regions doesn't work well in the designer, and so on...
My point is that I feel the whole thing is not mature. And Oracle seems to fix some issues with the release of JHeadStart. Sometimes you fix things by integrating JHeadStart stuff. JHeadStart is probably a great thing and a time saver if you don't do too much "exotic" screens and code, but does it means that JDeveloper+ADF and classic development becomes a second class citizen ?
This is sad, since the ADF framework and JDeveloper look good and are worth of a more widespread usage. But marketing and support are way way far behind other commercial and open-source solutions. When I see what is planned for JDeveloper 11 and I compare it to others (Java or not). When I look at the developers needs and expectations, I feel we are losing the race.
Maybe Oracle would like to hand over some of the framework to the community, but my feeling is that the community is already commited on other exciting and less constraining frameworks and tool, and this community may well work in a way that doesn't fit Oracle vision. Future will tell.
Cheers.