You're right. Slaughtering 20 junior school students and a crossing guard are not acceptable means to get your daughter from school to her dentist appointment on time.
It is however acceptable to kill 20 enemy combatants to rescue a dignitary from a besieged embassy.
While this helps get the point across, your example doesn't provide comparable events. The distinction is the weighing of the results of actions to the actions themselves. Is taking one life justified to save another? Is it measured by the greater good, such that the life that is saved will provide greater contributions than the one that is taken? Or is it measured in that all life is inherently to be treated as equal? And if all life is treated as inherently equal, is it a numbers game, taking 100 lives is justified so long as it saves 101 or more?
The phrase "The ends don't justify the means" is not hyperbole to discredit accomplishments, nor is it meant to remove the ends from the equation. It changes the equation to provide the necessity of weighing the ends against the methods rather than measuring the ends independent of the context used to accomplish them.
What is even more interesting is the opposite. The ends not justifying the means is typically used for weighing the actions necessary for a favorable outcome. If looked at from the opposite perspective, do the means justify the ends, it presents a more difficult scenario. If X people die, but only appropriate means were used (and others would have saved X people), should those that chose the appropriate actions be held responsible for the failure to provide a favorable outcome? Or should they be heralded for making the difficult decisions to only use morally acceptable practices?
Germans (or at least Nazis) would, at the time, have considered it to be entirely worth it. I believe that the intended context was that the dignitary was a member of the same organization/country as the people trying to rescue him/her, while the enemy combatants would be from a group opposed to that country. Your response, on the other hand, seems to suppose the existence of a third party (the dignitary, enemy combatants, and in addition a group which dislikes the dignitary and knows that he's a horrible person).
Although you are correct that I am assuming a third party, it is only in the role of judging the described actions as just or unjust that a third party is assumed...namely the gentle reader - a necessary assumption in any discussion of hypothetical ethical scenarios as the example of the school children illustrates.
There is a convention to these things - one doesn't assume that the dead school children were carriers of an incurable virus and that only their death prevented a deadly pandemic etc.
As I am sure you recognize, a patriotic motivation does not make an action right - even if your assumption that the rescuers and rescued share political affiliation might often lead one to assume your holding such a belief.
De-ontological ethics are hierarchically ordered. If one must choose between a lesser and greater good, one must choose the greater good. But, that doesn't make the action ontologically right.
I.e. stealing qua stealing is always wrong, but it is right to steal to feed a starving family, if stealing is necessary. However, stealing is still wrong.
>"it is right to steal to feed a starving family, if stealing is necessary"
This slips into what I called "semi-deontological" ethics which has the same issues as ontological ethics, i.e. that it is possible to provide a counter example for any generic ethical rule. Until you get down to an actual example one may invoke Nazi's, fatal pandemics, or lawyers tied to railroad tracks.
I don't think that while in war you weigth the enemies' lives (especially the combatants' ones) the same way you weigth those of your own. At least not until you approach anywhere close to a massacre of enemy combatants.
It is however acceptable to kill 20 enemy combatants to rescue a dignitary from a besieged embassy.