Re: XP and Pair Programming

"Robert C. Martin" <unclebob@xxxxxxxxxxxxxxxx> wrote in message
news:20jth1pm27mjdppt9b678mb662k3br35je@xxxxxxxxxx

> An accountable team has one, and only one, definition of the word
> "done". Towit: Passes all acceptance tests.

This is a very dangerous argument. I think I've tried to point
it out before.

Are the inputs of your tests serialisable? If they are,
I can simply write a single case statement that accepts
all possible inputs from your tests (since your write them
first, I know them all of course) and returns hard-coded
results expected by the corresponding tests.

This will pass your test suite. Do you accept that as
a valid soultion? If you do, it makes it so much easier
for the actual programmer. Heck, with a little bit of
effort it should be possible to write a program that
takes in a set of your tests and spews out that single
case statement to match it.

Note that it's also scales well. You, as a test writer
have to come up with a test case and find out what
the correct result is (I wonder how btw?) All I have
to do as implementor is to cut and paste your inputs
and expected outputs into the case statement.

No need for domain knoweledge or anothing fancy
like that - just cut and paste will do it.


.

Relevant Pages

  • Re: XP and Pair Programming
    ... >> An accountable team has one, and only one, definition of the word ... >a valid soultion? ... >for the actual programmer. ... Heck, ...
    (comp.object)
  • Re: XP and Pair Programming
    ... >>> An accountable team has one, and only one, definition of the word ... >>a valid soultion? ... >>for the actual programmer. ... Heck, ...
    (comp.object)