[Rainbowportal-devel] The argument for IBATIS
Brought to you by:
danijel_kecman,
manudea
From: Charles C. <cha...@gm...> - 2004-12-27 22:09:18
|
(I expect that anything being addeed to project code: MBUNIT, GENTILE needs a simliar justification and debate here and a concrete decision rather than any developer being able to add whatever they want and we all have to live with it in future coding sessions.) Rainbow Data Access code needs to be refactored no matter who you talk to for speed matters if nothing else so caching is not mandatory and to replace leaky code. While refactoring we should hopefully address the #1 need our RBN user base has. DB independence. What IBATIS does for us is (tada): once we make code with IBATIS to access and manipulate data in RB the access and manipulation interacts with an XML file -- we will call it coresqlserver.xml for this example. The XML file contains SQL server table info, SProc name or adhoc SQL, and parameter info. If someone comes up with better way to do the data access for SQLserver (different Sproc or table structure) they can make a new XML file coresqlserver_chaz.xml coresqlserver_v2.xml and then just change a config and without recompiling RB is doing it the new way. If someone makes: coremsaccess.xml coreoracle.xml for example these can be added into project and with a config change (no recompile) RB is using a different DB. If someone steramlines and speeds up the way Oracle or MSAcess is done they can produce a new XML file without touching compiled code and then swap out the new inmporved data access via a config change. |