On Mon, 14 Mar 2005 23:00:10 -0500, Sean Schofield <[EMAIL PROTECTED]> wrote: > Shale is definitely not off-topic IMO. I think MyFaces and Shale go > very nicely together. Plus every time you say the word "Shale" Craig > pops up here ;-) This is a cheap way to get him to answer our JSF > questions.
Look at that ... it worked again :-). > I suspect a lot us are going to be using Shale soon (I am about to use > it in a limited fashion in the next production release of our system.) > I notice a certain common set of issues that keep coming up on this > list and often times, the answer is "use shale." The short answer is that Shale is not functionally complete -- see the Wiki page for outstanding items. http://wiki.apache.org/struts/StrutsShale At best we could do some sort of milestone release, but I don't want to create any impressions that the API is close to being frozen yet (although ViewController seems pretty solid). But the best way to make the APIs solid is to get some real world feedback, so it may be that milestones are the way to go. Some other thoughts: * At TSSJS, I talked with folks on the Spring and Beehive projects. Both of them have abstractions to do what DialogController is focused on, and I want to experiment with both of them for this responsibility, to see if we can avoid having to reinvent that particular wheel. * Also, for the back end of XmlHttpRequest-sending components, we need to support APIs and message formats in addition to raw XML. Need to look at JSON and othe Ajax-ish things to see what we can harvest. * Implementations of support for file upload, client side validation support, and Tiles support have been promised. Regarding cooperation, that makes perfect sense ... among other things, I periodically make sure that Shale works with snapshots of MyFaces. One interesting question would be how to deal with components that want to be Shale-specific in order to use the remoting support. Maybe a sandbox subproject here at MyFaces, now that it's a TLP? > > sean > Craig PS: On the general topic of subprojects, I endorse what was suggested earlier about splitting the JSF implementation and component library, with separate milestones and deliverables. There's going to be lots of attention on both, and you won't want the delivery schedule for one of these negatively impacting the other, or have to worry about synchronizing them. If you go this way, the MyFaces PMC is still responsible for all releases for all subprojects. If you do things the way we did at Struts (which I would suggest), you'll also want to have universal committer access across subprojects to avoid potential barriers to increasing the communities around each subproject.
