My latest post to the agile-testing yahoo group:
Ben Simo asks how we can talk testing without talking philosophy.
If you will permit me, I will try to expand on that. As I see it, software testing is questioning the product, asking questions like "Is this Good (Enough)?"
To do that, you need to answer questions like "What does 'Good Software' mean?" "What does 'Good' mean?" and pretty soon you are asking "What does 'meaning' mean?"
The good news is that we can stop wayyyy short of that final question and still declare victory.
So we define good according to a value system - for example - we value this over that, because we value this outcome over that outcome -- applied philosophy. I am not the first person to notice this[1, 2].
So that is what geeks me out about testing. I suspect, that by philosophy, Lisa means to stay away from the arguments about "is testing critical inquiry?" or "is testing a repeatable process that should be automated all the time?"
This is very hard for some of us to do, as ideas have consequences, and our worldview will impact our work.
/AND/ here is what I can try (harder) to do in good conscience: When things get airy fairy, we can bring them down to THIS report, THIS web screen, or THIS server product, and it's constraints, and what test strategy we would develop.
I believe that's what we are being asked to do. I'm not excited about it, but I understand the concern and I'll try to keep things grounded in good faith.
Fair 'Nuff?
Regards,
--heusser
[1] http://www.satisfice.com/aboutjames.shtml
[2] http://www.paulgraham.com/philosophy.html
Thursday, April 09, 2009
Subscribe to:
Post Comments (Atom)
1 comment:
Most times when people decry philosophy, I think what they are really saying is "Can't we just fake it? No one cares whether we actually do a good job as long as we SEEM to be doing good."
When people say they don't need philosophy to test, I think they are really saying "we haven't yet been fired for incompetence by our bosses (who also know nothing about testing) so why worry?"
This is a critical issue with testing because testing is so easy to fake and there are such powerful incentives for faking it.
One of the reasons I no longer post on or monitor the agile-testing group is that the predominant philosophy of testing on that list is that programmers shouldn't have to think deeply about what testing means or what skills are required to do it well. Testing is just programming by another name, so the only thing worth learning is the newest test tool-- the newest way to make fake testing look flashy.
Post a Comment