If one wants to replace their halyards with messengers when they aren't in use, then the messenger should be attached (sewn) to the open end of the halyard and the halyard is then lowered completely where it can be stowed in a bag at the mast keeping the messengers attached or below if you want to separate them. This works for internal or external halyards.
It makes little sense to me to attach the messenger to the shackle and simply sky the halyard leaving the splice or knot exposed to the elements.
One caution; If using a very small diameter messenger, be sure to keep it under tension throughout the entire process or else the messenger may jump the sheave and become jammed between the sheave and cheek.
Knothead--You must have missed the notation that the exposed ends of the halyards are wrapped with Teflon Tape to guard against chafe on the sheave box edges and block UV. We've been following the convention of withdrawing unused halyards/lifts into the mast on our boats since the mid-60's (a trick we learned from an old, accomplished and much envied racer) and have never had a halyard/lift fail from other than chafe on the sides of a sheave box (no matter how industriously we smooth and polish the cheeks of the boxes). The objective is to get them out of the way/wind and protected until they are needed but to be able to put them into service quickly when needed. I don't want to be screwing around trying to reeve a halyard lead in the middle of the night in a heavy seaway when I can just pull a lazy halyard/lift down, snap it in place and be back in business.
I am in communication with the engineers at NE Ropes who assure me that the non-load bearing polyester cover quite adequately protects the load bearing Technora/Dyneema SK-75 core. Moreover, Dyneema is a highly crystalline UHMWPE fibre relatively immune to UV degredation hence its use in space science applications where UV exposure is not mitigated by an Ozone layer as we are here on mother earth. For the sake of comparison, we have uncovered 1/4" SK-75 runners that at are now 10 years old (that replaced similar 16 year old runners) with no failures despite constant solar exposure and not infrequent cyclical loading in the 2000 to 4000 lb range. The difference between these and our halyard that failed is the fact that the ends were spliced around thimbles.
While the post-mortum on our failed splice has not been completed, I suspect the failure arose because I allowed a supposedly knowledgeable rigger to make a splice without a thimble when I knew better. A tight eye splice around a shackle bale with high modulus fibers--no stretch--results in only the fibers on the outside of the curve of the eye being loaded and so bearing all of the load in the splice. These eventually fail with the load being transferred to the next outermost layer of fiber. The cycle repeats until the entire eye has failed.
In any case, once we have reeved a messenger, we shall end for end the line and reattach the halyard, around a thinble, but secured with a Buntline hitch rather than a splice and shall thereafter chop of a foot or so every few years and remake the hitch. Friends of ours now in Trinidad inform me they have applied the foregoing methodology with their, now, 20 year old Dyneema French issue halyards and have suffered no failures thus far.
And, we shall continue withdrawing our unused halyards and lifts into the mast as a matter of course. Different ships, different long splices eh?