I have recently seen a theme come up a few times. Worth restating general principles.
Some splinters are made as quantitative shape-bash bids in some sequences. In other words, "I have a stiff and slam interest" is all that we know.
That said, GP suggests that LTTC be available, at a minimum, if available.
Thus, consider a sequence where spades is (or will be by the splinter) agreed.
A 4♥ splinter sends that message, but the range should be rather tight, because partner must decide now, on this information alone, whether to resign to game only or venture into the five-level, often without a good safety net, on a slam try.
A 4♦ splinter, however, can be more flexible. This is because partner has three options. First, he could "Hell No!" your try and resign to 4♠. Second, he could giddily charge forth into the five-level. Third, and critical to the point, is that he can send it back at you with a 4♥ Last Train bid. The ability for partner to hedge here allows you to be more flexible with your two-under splinters, showing either a good sound splinter (will accept a LTTC bid) or a weakish splinter (will decline).
Three-under splinters (4♣ here) have even more room with which to work and can, therefore, be even more flexible. One might want to devise some techniques here to distinguish, for example, a 4♦ reply to a 4♣ splinter, a 4♥ reply, and, as to the 4♦ reply, the meaning of 4♥ by the person who splinters.
There are even four-under splinters, where LTTC and Serious 3NT are both available. 1♥-P-3♠ is a classic example. Opener can start cuebidding with or without a Serious 3NT if he wants.
The point, ultimately, is that there is not a need to define all splinters equally, or to treat them equally. Recognize this space principle, and recognize partner's recognition of the space principle.