Is the Microsoft UI Automation Framework window processing for Java SWT, C ++, Delphi, or should I use OCR?

I have been working on AutoIT and window automation for over 03 years, but mostly on WinForms. AutoIT does not work in WPF or Java SWT UI. I used the White Framework a bit. I want to write a general application for testing and automation of the 3rd party interface, to which I pass xml with a list of window title, text window, interface and user interface text (on the right or left) - just like AutoIT-click control, and it performs an action.

My requirement is that it should work on all Windows desktop applications (regardless of whether they are written in C, C ++, Net Winforms, .Net WPF, Silver-Light, Java, Delphi, Qt or whatever something else). The end user will be our operational team or testing team (who will not be able to program). What approach should I take?

• Does Microsoft UI Automation support the infrastructure of all types of Windows user interface automation?
• Is the white structure better than Microsoft UI Automation (I checked - the white inside also uses Rhino)?
• Should I write my own library that uses OCR to detect controls (same as user-to-user)? What disadvantages will OCR use?

Relations Akshay Mishra