Hi folks.
I spent the day examining the way three separate commercial applications (that perform similar functions) expose a COM interface. I saw three completely different approaches:
I'm trying to decide on an approach for my next application. I was wondering what thoughts you may have on these approaches - either from an implementation perspective or from a 3rd party perspective (ie. one who might be wanting to automate one of these applications).
I spent the day examining the way three separate commercial applications (that perform similar functions) expose a COM interface. I saw three completely different approaches:
- OOP structure - A complete heirchy of collections and objects representing the database (and limited methods/properties for the application itself)
- GUI encapsulation + OOP - methods/properties to fully automate/control the application (ie. just about every menu option) with a heirchy of collections/objects representing the database
- Flat access/interface - a single interface containing methods to retrieve/alter any field in the database using named fields (on the active record). No OOP structure at all, limited options to automate/control the application.
I'm trying to decide on an approach for my next application. I was wondering what thoughts you may have on these approaches - either from an implementation perspective or from a 3rd party perspective (ie. one who might be wanting to automate one of these applications).
Comment