Re: Re: Large Java classes wanted

J Arrizza <[email protected]> Fri, 26 Feb 2010 22:34:45 -0800
Newsgroups gmane.comp.programming.refactoring
Message-ID <[email protected]>
> If your extract class algorithms can't work with the "reverse
reverse engineering" case, it probably won't work in the more general
case. At the very least, the "reverse reverse engineering" case
should provide some test cases while continuing the search
for human-generated test cases.

Yup. If you don't set up hand-crafted ATs (the "reverse reverse engineering"
cases), then you will drive yourself crazy if you jump directly to real-life
test cases. If they fail, you have no core set of test cases you know pass
and so you're left with thousands of lines of code that may or may not be ok
or not ok or...

Same sort of strategy applies when testing things like AI apps
or probability-based apps (e.g. spam detectors). You need a suite of
specially designed hand-crafted ATs that check the basic functionality of a
spam detector, but then there's nothing like real spam to check if it will
actually work. You can test all you want with rigged up test cases, but you
still need real-life human-generated test cases to double-check the
acceptance tests aren't inadvertently passing with false positives.

In short, you need both.


[Non-text portions of this message have been removed]



------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/refactoring/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/refactoring/join
    (Yahoo! ID required)

<*> To change settings via email:
    [email protected] 
    [email protected]

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/