Re: Roadmap Updated

Hi Gary,




I feel that you've grokked Theories more than us (well, me at least!) and
have some strong ideas about it, but you're not giving much away ;-)  is
that to get our ideas out of us without us being biased by first hearing
your thoughts?
 
Yes.
 
<meta>I have a lot of experience with folks not questioning my ideas. It's a
problem
for me. Maybe I project too much certainty. Maybe my ideas are just good.
:-)
</meta>  
 
 If not, it would be great to hear what exactly theory means to you, and
some ideas of what you think this means to NUnit v3 - maybe with some
examples? I've only read the first paper on theories that you linked to, you
did send links to some more resources but I haven't had time to read them
yet (which may be part of my lack of understanding ;-) )
  
Take a look at the Farenheit/Celsius example in my last post. Does that
help? Theories
are... theories - just like in science. We need data to "prove" them - in
quotes because
you really can't prove anything by examples, just get more certain.
 
Before the Theories papers, we had lots of ways to generate data, including
RowTest
for example. David came up with a concept that wraps it all up in a very
clear way.
 
In theory <g> you could demonstrate a Theory without any data points. For
example,
by analyzing the code. You can also use any of the techniques for expressing
or
generating data points without having a Theory - "I think my code may work" 
doesn't count.
 
In developing, the Theory comes first, e.g:
 
"Any square root, multiplied by itself, gives the same number."
 
The data points come by asking oneself "how can I demonstrate this?"
 
If you just start out with a method to test and throw data at it, you're
doing data-driven testing, but not using a Theory. There's a place
for that, but I think the notion of Theories could change the flow
of development in much the same way that TDD has done.
 
What do you think of the theory implementations in xUnit.net/JUnit? What are
you planning to be different in NUnit?
  
I'm keeping my mind open. For obvious reasons, JUnit's implementation
follows
David's papers, so I tend to use that as a benchmark. I've only looked very
briefly at the xunit implementation. <meta>I tend not to look at other
folk's
implementations till I've taken a shot at it myself first. That's not a
moral
stance, just my personality.</meta> 
 
I suggested that Kelly move the Theory discussion to NUnitV3 because
that already seems to have the people with interest in futuristic stuff. Of
course,  if everyone wants to talk about it here, that's OK too but I hope
those with an interest in 3.0 will join nunitv3 as well. 
 
I'm about ready to write about my own thoughts on this too - as soon
as Agile Open Northwest is over.
 
 
Charlie
 
Cheers,
Gary


  _____  

She said what? About who? Shameful celebrity quotes on Search Star!
<http://www.msnsearchstar.com>

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

_______________________________________________
nunit-developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/nunit-developer