This post originated from an RSS feed registered with .NET Buzz
by Udi Dahan.
Original Post: Focusing on Value
Feed Title: Udi Dahan - The Software Simplist
Feed URL: http://feeds.feedburner.com/UdiDahan-TheSoftwareSimplist
Feed Description: I am a software simplist. I make this beast of architecting, analysing, designing, developing, testing, managing, deploying software systems simple.
This blog is about how I do it.
I almost skipped over the required "back from..." post in my rush to put out all the content building up in my head after the software architecture workshop in Cortina - so this will be that post.
Anyway, after reading "Innovation Stifled by Dogma", I was reminded of one of the session we had at Cortina - about IOC and DI. What reminded me was our attempt to focus on the Value produced by these practices, rather than getting too bogged down in techniques or philosophy.
The goals we identified that drive us to use DI, or at its more basic level, dependency on interfaces instead of implementation (DI or service locater being refinements), were simple, and important:
Testability and Runtime flexibility.
By introducing interfaces, we can test classes in isolation, creating tests that will give us fewer false negatives - tests that fail even though the class under test behaved alright. Higher testability often leads to fewer defects.
As for runtime flexibility, the ability to vary implementation without affecting system structure is interesting. Beyond the ability to fix bugs in a given implementation without adversely affecting other parts of the system (a worthy goal in and of itself) I hadn't really thought this one through. The idea of so-drastically changing an implementation, not with mocks/stubs, but actually changing one with something really different, seems a bit out of touch with reality to me - but maybe its just me.
Anyway, from that group of IOC and DI practitioners I gained much deeper understanding about "why" which has really helped me change the "how" of my work - but I'll save that for a later post.
Previous Topic
