[SciPy-Dev] deprecations - things to remove before 0.8.0

[Stéfan van der Walt]

On 27 May 2010 17:28, Robert Kern <robert.kern at gmail.com> wrote:
> The reason to *follow* a rule is not the same reason for *instating*
> the rule. We should follow the rules that we have agreed to because we
> should make good on our promises. Otherwise, we might as well not make
> those promises and make just make deprecation/removal decisions on a
> case-by-case basis all the time.

Sure, that makes sense except that the rule never envisioned this
situation: deprecation of a broken function.  What do we do: raise a
deprecationwarning and then proceed to let the broken code run?
David's argument about monkeypatching also doesn't make much sense to
me: if a user monkey-patched ppimport, how will deprecation help?

If we deprecate, we have to fix the function to at least return a valid result.

Regards
Stéfan