First, R: Touch doesn't say that. R: Touch leaves it vague if you touch first and then finish the spell, do everything else and finish with a touch, touch the whole time, etc. Second, even with touching after triggering it's fine: the triggering ends with the moment of impact (last bit of triggering sequence), and the weapon must be touching the target a moment after the moment of impact - it is physically impossible for this not to be the case.
To give you the benefit of the doubt, I assumed, that by the above you put the thumb into a position where it cannot be broken if you strike. This isn't the case with the posture in the above picture: so I excluded such postures. You tell me that you wish to include them now: fine. But this posture makes striking "likely to break your thumb" especially, if the fist hit an upward or sideways moving shield, vambrace or weapon.
First, if you want your thumb where it can't be broken when striking with a fist, cut off your thumb. Second, wrapping the thumb at least roughly like that puts it in one of the most protected spots it can be in when hitting a hard surface with your fist. This is one reason why professional fighters do it. You can also see this in skiing, where the safest position for your thumb when striking a hard surface with your fist is to have it roughly where shown in that photo. Moving it to spots other than shown in the photos above tends to drastically increase the risk of a thumb injury. But maybe professional martial artists don't know how to use their fists and maybe real-world stats on the large number of skiing crashes resulting or not resulting in broken thumbs are all wrong? Should I trust the professionals at what they do or what you say about what they do? Should I trust the real-world stats on thumb injuries or what you say they should be?
Well, that's not selective reading. It's an ArM5 p.85 erratum anticipated in an obscure spot on RoP:M p.125.
No need to anticipate an erratum when there is no error to correct:
ArM5 85 has the following on magic resistance: Given (ArM5 98) do they have to penetrate resistance because they are effects not spells and hence not excluded by the exception in ArM5 85 above? If this is just unfortunate wording does that warrant an errata? If this distinction is indeed relevant, should this be made explicit in ArM5 98 and hence warrant an errata either way? The topic came up in a discussion involving Touch range effects triggered during combat, potentially via InMe based…
Item effects and being powers are designed as spells. For invested items the core book even says they "mimic the power of spells" (ArM5 p.95-96). While effects aren't spells (so not cast the same way, but created in the lab), "an effect is like a spell, and must be fully defined, as a spell is" and "you must design the effect carefully, as if inventing a new spell" and "the effect is based on the spell guidelines" (ArM5 p.98). I think those are sufficient to say you follow the normal Hermetic guidelines for building spells, and Range (which includes Personal) is a part of that (ArM5 p.111).
Every core Hermetic Range says "spell," "magus," or both. Why aren't you objecting to all the other Ranges (as well as other stuff like Durations)? As for Personal in particular, Personal is defined based on a spell. Why are you excepting "spell" in one case for Personal and not excepting it at the same time in the other case for Personal? That's selective/mis-reading.