Re: Config class

excuse me for using your thread, but i have some things to point out about this class, so maybe it`s a good place for it: - why in all get/set methods key address is splitted in two parts (path+name)? wouldn`t it be easier for developers to use one full path? class should split this internally imho - i find $data array construction quite confusing. using flat array to reflect something that isn`t flat at all (think of xml as a base container, not ini files) isnt good idea in my opinion. i know that its class inside structure and it probably doesn`t matter for end user, but as i was thought - code should always reflect reality, not create one. my proposition is to use nested arrays to store configuration data - just to make things clearer. or maybe performance was the most important thing when you decided to use flat arrays? of course additional ticks for going through array would be needed. - does anybody (except me :)) use wddx container? if not than it wouldn`t be bad idea to withdraw it from cvs. my goal was to create container that supports writing and block nesting at the same time (no matter how bad it looked in files) - if you are going to implement this features generally (or did i get this thread wrong?) - there is no need to leave wddx container. when i was writing wddx container there were no signs of full feature xml container in near future. btw - don`t get mad if my questions were answered somewhere in deep past, i wasn`t here when you were making first drafts, so just give me some hints why is that made in this or other way :) rash