I recall my instructor preaching against diving for the glideslope, stating
that dropping at over 1,000 fpm at a low altitude and in IMC could be
problematic.

And he's right - it CAN be problematic. It demands more of the pilot.
Set up that descent and divert attention for a bit longer than you
planned, and you can be in for a once-in-a-lifetime experience - the
kind that comes right at the end. But sometimes it's necessary to get
the job done. So how do you know when it's appropriate? Believe it or
not, there is an answer.

It's appropriate when you can see in advance that you will have to do
it due to factors beyond your control. In other words, it's OK to do
this to fix a bad vector - but not your own mistake. Why? Because if
you already made a mistake bad enough to put yourself in this position,
what makes you think you won't make another that bad? A radical
maneuver that requires better-than-average skill to pull off is a bad
idea if you're using it to fix a mistake caused by your own
worse-than-average performance just minutes or seconds ago.

On the other hand, when you have to do it to fix the mistake of someone
else, one you saw coming as he was making it, it's not a big deal.
You're starting out ahead, not behind.

As the more experienced pilots in this thread pointed out, apparently this
is a viable tactic, but certainly one that develops with experience.

I teach it as part of the initial instrument rating - because this kind
of problem is so common. I will actually create bad vectors for the
student to fly, and teach him how to deal with them. Given what I've
seen at Houston Approach, it's just common sense - he will be dealing
with them sooner rather than later. But that comes AFTER the basic
approach is mastered, and I never allow the student to use these
techniques to fix his own mistakes.

Michael