Re: Doubt about ADF Faces & Trinidad

[Frank Nimphius]

Stephen, some comments to your concerns, especially in regards to the new JSF component library Oracle builds * it took forever to negotiate the legals of the first adf faces drop, then it took another   eternity until Trinidad 1.0 and again it took a series of minor releases until Trinidad was   mostly stable Yeah, it took a while to get Trinidad out of incubation, but I think that good stuff needs time to mature. The Apache folks take good care that no rubbish appears on their website and that all projects are with a healthy community driving them. Frankly -  I am quite thankful for this. * Oracle employees still do most of the work Because we have an interest in this component set because it replaces ADF Faces in JDeveloper 11. In this regard, Oracle is a customer of Trinidad and happy to contribute as good as we can * many people enjoy working on experimental stuff or whatever else they like, all while   bugs that affect basic functionality pile up in Jira I think that all software products suffer from that, especially if bug fixing is on a voluntary basis. In commercial software the author of a class fixes it if it has a defect. The open source model however is different and anybody in the community can provide the fix * documentation is, hm, let's say minimal. There isn't even some kind of visual index to the   components. No small usage example at each component's tag docs.  That is even a step   backwards from ADF Faces. I think that ADF Faces was just a headstart for Trinidad not to start from scratch. Trinidad is a component set on its own and while Oracle has professional doc writers building the collateral, Trinidad is dependent on volunteers and their contributors to make sure the documentation improves with the product. I also think that filing bugs on documentation is a first step into the right direction * Now instead of working on Trinidad 2.0 Oracle decides to do yet another component library.   Discussions about the initial drop are undergoing, so we might see a 1.0 release in 2010.   By then it will be largely incompatible to the initial drop with all the subtle difference   being poorly documented. I think you misunderstood Oracle's role in Trinidad. We contribute to Trinidad and meanwhile have our own projects that we need to drive for internal and external customers. we also spend developer resources on our own JDeveloper IDE, EJB, JPA, ADF binding etc. The ADF Faces rich client component project is a step towards simplifying Ajax development. Its a pioneering work that - once this is released to open source - again becomes a major headstart for the community "[...]I recently spent so much time debugging and trying to beat skinning into form, work around incompatibilities between Trinidad and Seam, etc. I ended up writing a couple of custom components for the core functionality.All in all the effort would have been smaller if I just went with JSP from the start[...] Sure, the problem however is neither Trinidad nor Seam but the combination of the two. It is not uncommon that different component sets don't integrate well with each other. In the Web Services space it required a WS interoperability group to solve incompatibility issues. I think the same is needed for JavaServer Faces, especially because we know that Ajax in JSF wont make things better automatically (I suspect this to become worse instead). So the question is if incompatibility is a reason to stick with JSP or if it should be a motivation to look for how this can be changed to the better. Frank Stephen Friedrich wrote: Mark Millman wrote: The quality of Trinidad it a testament to how Open Source communities work best. Well, ... I don't really agree, but maybe I am not involved deeply enough in the community to judge. Without wanting to offend anyone personally - it seems to me that * it took forever to negotiate the legals of the first adf faces drop, then it took another   eternity until Trinidad 1.0 and again it took a series of minor releases until Trinidad was   mostly stable * Oracle employees still do most of the work * many people enjoy working on experimental stuff or whatever else they like, all while   bugs that affect basic functionality pile up in Jira * documentation is, hm, let's say minimal. There isn't even some kind of visual index to the   components. No small usage example at each component's tag docs.   That is even a step backwards from ADF Faces. * Now instead of working on Trinidad 2.0 Oracle decides to do yet another component library.   Discussions about the initial drop are undergoing, so we might see a 1.0 release in 2010.   By then it will be largely incompatible to the initial drop with all the subtle difference   being poorly documented. I like JSF in principle, but it is still lacking a professional, complete, standard compliant and compatible components set. Sorry if I sound harsh, but I recently spent so much time debugging and trying to beat skinning into form, work around incompatibilities between Trinidad and Seam, etc. I ended up writing a couple of custom components for the core functionality. All in all the effort would have been smaller if I just went with JSP from the start. -- ________________________________ Frank Nimphius Principal Product Manager Application Development Tools Oracle Corporation mail: [EMAIL PROTECTED] phone:+49 2058 782481