On 4/23/06, Martin Cooper <[EMAIL PROTECTED]> wrote: > > On 4/23/06, Don Brown <[EMAIL PROTECTED]> wrote: > > > > I would think it would be Tiles' responsibility to support deployment > > with Struts. In that view, Tiles would be its own project, yet part of > > it would depend on struts-action.jar to provide the Struts hooks. In > > the same way, they would provide Struts Action 2 hooks, if necessary. > > > This is a tough one - for me, at least. Should an independent Tiles have > glue code for Struts, or should it be Struts' responsibility to provide > the > glue? If we look at Velocity, the glue is over there, but we've also > talked > about it coming here. If we look at Validator, the glue is here. If we > look > at Chain, the servlet and portlet glue is over there.
I've always thought it would be as Martin describes for Validator -- standalone Tiles would really be standalone, and each framework that wanted to utilize it would provide it's own glue to some particular version(s) of Standalone Tiles. That is what Shale already does with the standalone Tiles stuff -- for example, Shale integrates tiles into the standard JSF navigation mechanism. That's not something that it seems reasonable to impose on a "standalone" project. My preference, at least at this time, would be for the Struts / Tiles glue > to stay here. That will help Standalone Tiles stand on its own feet, and > especially help with the notion that it's now independent of Struts. Any > perception - real or otherwise - that Tiles is tied to Struts will be > detrimental to Tiles as an independent library. +1 -- > Martin Cooper Craig
