If you use the target message, you must explicitly manage voice allocation yourself. This can work well in the case of an additive synth, where each harmonic has its own poly~ voice and you never ask for more voices than poly~ has.
When you don't know how many voices will be needed and voice-stealing might be needed, the note or midinote message are easier to work with.
The difference is clearly seen in the example patch "stockhausen-studie-II.maxpat" located in the examples folder. If you unlock the patch and double-click on "StII-synthesis", you'll see a poly~ playing the notes of the score. Double-click the poly~ to see another poly~ inside creating the harmonics of the sounds, using the target message.
Voice stealing is useful when a synth receives more notes than it can process. Voice stealing ensures you hear the most recent notes, while killing the older notes. This strategy works well musically.
Check out the granular example in the examples folder. The poly~ in rgrano.maxpat has 16 voices, but you could easily ask it to play too many grains if you shorten the metro time and/or extend the duration of the grains.
A more typical case would be if you made a keyboard-controlled synth and played big clusters of notes with your arm. By killing older notes and letting the new ones play, you hear a more accurate representation of musical intent.
one might call that a good compromise between processing power and what would be musically ideal (e.g. 88 voices for a piano sound) but the truth is that it has been invented to impose a synth having more voices than it actually has.
of course you are right that you usually do not hear a missing release of a typical synthesizer voice under the condition, that it is only 1 note among a total 16 notes currently running. but try the same with a harp ...
stealing might give you like 20% more felt voices, but sooner or later there _is an end, and you still have to care about how you play the virtual instrument when you want to avoid noticeable cuts, just a bit later.
i believe using the "note" system of poly~ will not teach people what they need to know when they want to build a string instrument style (or "mute groups" style) voice stealing mode next time.
regarding poly~, isn´t it pretty easy to implement voice stealing using the target message, too? what was first btw? did the object eventually had only the note thing in the beginning?
As I recall, a number of older analog synths had a variety of approaches to voice stealing. There are different strategies which might be appropriate depending on the musical context. High-note priority, for example, steals notes other than the highest pitch currently sounding - in some tonal contexts it makes more sense to preserve the top note as it would likely be the melody. Conversely, low-note priority preserves the bass note, assuming it is helping to establish the tonality.