Re: Re: Package thoughts: XML_RSS_Publisher

Hello Lukas, From Monday, January 17, 2005, 11:06:46 PM, you wrote: >> >> The main processing engine would be use the XML_Serializer - that's what I'm >> using at the moment - and why reinvent the wheel? >> >> Currently I'm building a hash of categories - either statically or via a >> database. >> The Database side of things uses a subclass of DB_DataObject and uses >> defined links to pull extra items - if necessary. >> >> As nothing existed within PEAR already - I thought I'd wrap it up and define >> an interface. LS> I wouldnt force the use of any storage layer at all. LS> To be honest in my opinion it seems to be like this is more of a sample LS> of how to use PEAR. Which is obviously not a bad thing! LS> However it seems to me like its going to end up being just a few lines LS> of glue code with probably just as much config code. Turning this into a LS> package itself is likely to be less flexibile or create too much overhead. +1 Tutorial on DevShed and in XML_Serializer documentation would be enough to find information for everyone. I use XML_Serializer for RSS too. Also "PEAR tutorials" chapter in documentation describing solutions for typical problems and showing typical uses would be nice. User comments will have to be transformed to that part one day. You can also include links to DevShed and other sites with PEAR information on Documentation tab. t --