Hi, Lets say I have this webapp based on Struts and Tiles. One page is for displaying the contents of a database table. So, of course, I created this small custom taglib to encapsulate the mechanics required for iterating through a ResultSet and displaying its records in a table.
Now, new features have to be added like paging and editing. In an effort to keep the design clean, I wish to do some of the processing before the tags get evaluated like creating a beans collection with a subset of the ResultSet. I have to decide where this is going to happen... ...in a bean, called from the JSP page with a reference to the pageContext or directly in the page as a scriptlet; ...in a Struts action serving as a front to the JSP page; ...in a descendant of org.apache.struts.tiles.Controller that is specified with the controllerClass attribute of a Tiles definition. ...in some other place? Please tell me. I leave the first option out. I would like to keep the page simple and free of Java code whenever possible. It could also be said that this stuff doesn't belong in VIEW components. Then I look for documented benefits of using the Tiles solution. I can't find any. I dig a while in Tiles'src and try to understand its inner workings... Finaly, I come to the conclusion it offers me no advantages over the Struts Action. Have I missed something? Perhaps I should keep the Tiles solution for times where I want to have access to Tiles definitions programmaticaly. What do you think? Etienne
