You could also argue that knowledge of how to use swords is very useful in defending against swords, even if you're currently wielding a mace. Of course, if you and your opponent are both using maces then it's hard to say exactly how skill with a sword would help.
There's another reason to play it as written...it's easier on the player, less bookkeeping. It can sacrifice realism in some situations, but nobody in playtest seemed to notice. Otherwise, it seems to me, it forces the player to keep track of one "sub-DP" for every weapon group specialization, as well as a a sub-DP for every mastered weapon. I was afraid that would make it something of an accountant's game?
And yet, if a GM were to ask me to keep track like that, I'd have no problem with it. What I'm saying is I went with the simplified version for the majority, because adding in the extra conditions and complexity is easier than starting with a lot of contingencies in print and scaling it back later.
Edit: I took out the words "there only one reason to play it as written", because as Randall has shown, that's not entirely accurate. We can always find ways to justify the rules, although it gets into dangerous territory for debate when we seek to justify them in terms of what's "realistic". Those debates I personally try to avoid, because I suck at them. LOL