While on my way home tonight, one of the thoughts in my head was (paraphrasing): Does the relationship map in Smallville need to be the fundamental mechanic, or could the concept be lifted and used simply as a brainstorming convenience? More generally, when is any concept worth enshrining in a rule, instead of being a guideline or a process used to provide the campaign structure.
An aside, I'm unlikely to run Smallville, but I could totally see borrowing the relationship map to direct character and aspect creation in a FATE game. In fact, Values or Beliefs sound like ideal forms of a Named Aspect, to tie it back to my earlier brainstorming.
There may be a point of diminishing returns, but how about if a Dresden Files character's aspects were all a bit more structured:
1. High Concept
3. Core Belief
4. A relationship you have with an NPC
5. A relationship you have with another PC
6. A relationship with a place.
7. Wildcard-- take whatever you want.
Tying background phases to specific aspects isn't necessary in this. You might decide the most important NPC is a parent, or it might be the woman you met last week. But everyone should care about or have a relationship with someone. Everyone has SOME core beliefs. That's doubly important for Wizards, since you can't do something with magic that you don't believe in.
Digression over, back to the original question. The advantage of enshrining a concept into rules is that it provides more structure than simply using it as a process or player activity. Both processes and rules provide structure -- the bones that the GM and players can put some meat and flesh onto. As aways, with reasonable and skilled roleplayers, structure is less necessary -- all things become possible. :P Exhibit A would be Amber Diceless. Amber has very few hard and fast rules; it really comes down to the GM's internal decision making process.
So one possibility is to pilfer games for their best concepts, but not enshrine those concepts as rules. Steal relationship maps and use them for brainstorming. Don't bother with mechanics, just do it as part of character generation, draw it out, and have everyone incorporates its ideas unconsciously. As I think I said in an earlier post today, one of the frustrations for Houses of the Blooded was that I never managed to feel 'connected' to the setting for my character. We had a couple of small relationships worked out, but only in broad (one might say bland) terms. A detailed Houses of the Blooded relationship map would have certainly helped draw my character in more.
At the same time, if the concept is really important (to the GM, the setting, the theme or the tropes), "defining it with game mechanics" is reasonable. Defining it in mechanics, even if those mechanics are relatively simple, makes it easy for GMs and players to agree on a structure for that part of play. If things are always in the GM's head, things can sometimes feel arbitrary. Sometimes being able to know where you stand (we need 3 more successes to win, and we've got 2 rounds to get them) is a real boon.
Thus, I don't have a problem is a Leverage game defines variants of "Skill Challenges" to encourage specific tropes-- in fact, I think there's still room for a book of 1001 skill challenges and it can be system agnostic -- a huge number of games use (player total) vs difficulty number mechanics, after all. This, a Leverage "timed action" is something where the mechanics are easy to port to other systems.
I think the ideal situation probably starts with a simple, consistent core mechanic (like the Cortex or FATE die mechanic) but leaving enough room to build interesting mini-games on top of it. Ideally, the mini-games should designed to be reasonably transplanted or modified for the campaign's usage. Smallville's mechanics don't quite that bill for me. Radical departures in rules mean that you can't borrow the rules as much as you can borrow the underlying concepts. It suggests that when designing or describing rules, start with the concepts, and spell those out first.