James G. Sack (jim) wrote:
Well, assuming it's a deterministic system to begin with, and assuming
that it's possible to calibrate/validate it with human assistance (for
input X, correct output is Y), then it seems that should at least tell
you whether the transform function ever changes.
True. And I developed much of the system before the real-time data was
available by writing code that stuffed fake records into the database.
It's helpful. It's just not helpful enough that you can rely on tests
catching problems.
I appreciate that measuring and tolerances can be tricky, but in
principle, for a given input, you want a specific output, no?
At a given point in time, yes, that's what you want. At least in my system.
The problem is it doesn't test what happens when the feed changes, or
when you have a race condition, or etc etc etc.
Helpful, but far from being able to do "test-driven development."
--
Darren New / San Diego, CA, USA (PST)
"That's pretty. Where's that?"
"It's the Age of Channelwood."
"We should go there on vacation some time."
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
