On coincidence I found your Scrum pictures on flickr, read a line or two here and thought you might be a good person to talk with. I am currently writing my master thesis on Scrum and I am looking for creative people working in Scrums being interested in taking part in. Would be nice hearing from you, if you are

With best regards form Trondheim,

Christoph

@beza1e1

Your circular dependency could be broken by versioning.

True, but as I say in the post neither Clojure nor Java have versioned packaes so I would have to either create a hacky version (i.e. changing namespaces to have their versions in their name) or implement my own versioning system.

@RobertJ

I’m not sure you got the idea. It’s an analogy:

# S: (adj) analogous, correspondent (similar or equivalent in some respects though otherwise dissimilar) “brains and computers are often considered analogous”; “salmon roe is marketed as analogous to caviar”

( http://wordnetweb.princeton.edu/perl/webwn?s=analogous )

@ifatree

Yes, in the end that’s the approach I’m taking. Fato itself is tested by a simple application that uses it, if this app’s build pass then fato’s build is ok.

@Titus

Interesting. Why wouldn’t a test (forget about TDD) e a proof of *any* kind? I’d like to understand your comment better.

Thanks all.

Another way to state Godel’s theorem is, I think, that within any sufficiently interesting formal system it is impossible to prove or disprove certain interesting theorems. Were your programming language a formal system, you could probably draw a stronger analogy; as it is, your language and platform are implementations of some (not necessarily well-specified) system and so long before you get to formal correctness you have to worry about implementation details.

At least, that’s what I think.

mock up a static app (that won’t change much) to exhaustively work the testing framework, and make sure that one always works before you assume your framework changes are green.

otherwise you’ve just promoted your test app to production status without leaving a test copy behind. struct-quack was essentially manually testing fato. now that you want fato to test struct-quack, you need to pull out the pieces that can test fato directly and make a third app underneath.

i hope this concept maps to clojure. i’ve seen people with similar questions lately in every language trying to get their heads around testing. it’s not circular unless you make it circular (which is bad).