Arne said:
Lew said:
It is such an overwhelmingly bad idea to throw staff into the middle of a
programming team in the middle of a project, with such consistently negative
results when tried as has been objectively verified for decades, that Brooks
is not the only one to offer the advice to avoid it, nor to substantiate that
advice with evidence.
Actually Brooke [sic] does not provide empirical evidence just a made up
example.
And call the law for "Oversimplifying outrageously".
Well, all right. Others have provided evidence. See McConnell's /Rapid Development/,
for example.
There's an exception to every rule, except this one.
While Brooks called it an outrageous oversimplification, it turns out to be neither
outrageous nor overly simplified.
It has been observed consistently in the decades since /Mythical Man Month/'s publication
that adding people late to a project tends to hurt more than it helps.
Is this an essential characteristic of adding people late? Of course not, but in practice any
project that throws staff into a project that's in emergency is not hep to the best practices
of software project management, thus their emergency in the first place. For that most
common scenario, the likelihood of such benighted management suddenly knowing how to
integrate late staff additions is near enough to nil to strongly militate against betting in its
favor.
Theorize all you will, the practice bears out the principle.