@nasKo I won't comment on everything on your Dec. 5th post, you guys seems well aware of what you want to do and the reasoning behind timing issue (e.g. UI update when all or most feature are present to avoid multiple redo).
However, if you allow me this brief rant, I'd like to comment on the leveling system… because it's sort of a hobby/passion of mine I don't believe I know everything, but I hope this will help in a very though job that either define the work or, to employ words of others, "is too generic".
While the issue you raise (gathering sort of general XP to spend on possibly unrelated skills) does indeed seem a bit unrealistic. It does however present some advantages by it's relative simplicity and flexibility. Your system as is, is fairly OK and I'll explain why further. Some variation of this system are completely ridiculous, in theory, in most MMO (at least those I played before stopping entirely about a decade ago) and a lot of RPG : if you buy an skill/talent or whatnot in the tree when you are level 1 or when you are mid to high level, the same skill cost exponentially more experience. The linearity of your "skill tree" somewhat mitigate this issue.
A model like TES (any of those I played since TES III) may seem like a way to fix the previous issue and afford some degree of realism. However, there are several possible issues to these kind of system.
Issue 1 - The link between difficulty and leveling-up
This issue won't probably be one for PZ, so I'll go over this briefly.
Most if not all RPG, including TES, have to balance increasing difficulty against the player progress to make sure that (a) the player is neither bored nor overly frustrated for most of the game (b) having a fair and fun game. As such, it can create a strong pressure to mostly specialize in one or a tight group of complementary skills when the difficulty is strongly tied to the skill levels. This was especially the case with TES III and IV where you needed to level-up 10 of your major skills to earn a level (there was a workaround though, but it was counter-intuitive and broke the immersion factor). This has somewhat been mitigated in TES V with what seemingly translate in any skill's XP being added to the next level progress (I'll come back to this idea of having multiple XP values).
This issue is somewhat shared as described earlier for MMO/RPG if the player select a lot of low level skill while at higher level. However, player might pick more naturally the higher level skill he just unlocked then older ones he/she had access to for a while, thus reaffirming the simplicity and usefulness of a point based system as opposed to a grow as you play system.
Issue 2 - It becomes prohibitively boring to level secondary skills
The leveling and difficulty issue aside, it may be useful to a specialist to have some secondary skills that have a reasonable efficiency. In TES, it might be a warrior with some healing (Restoration and/or Alchemy), but to keep the effect relevant with it's pool of HP it would need to be of a decent level. That would mean in practice that you would end-up casting magic spell idly and artificially to raise the skill beyond it's natural level (i.e. without any power leveling). This can and has been mitigated by trainers and skills books, however it remains a marginal solution in it's (TES) application.
In PZ terms, this question would be relevant with similar roles, especially in single player or small group coop. A looter (i.e. adept with either or both weapon and stealth) might want to pick-up some first aid (for broken glass wounds only, of course) to increase his or hers likeliness to survive solo trips. It would make little sense that a player had to injure himself or herself to level-up first aid if it did not happen to be otherwise a lot in use (thankfully).
Of course, this is an issue only when the only way to level-up a skill is through it's use. If there are reasonable alternatives ways, such as the previously mentioned trainers, to increase a skill without it's normal use, it mitigates or eliminate the issue. It then becomes the though eternal questions of balance and adaptation.
The balance goes without saying, especially for PZ where there will be a blurry line between a strong urge to avoid ways to safely and quickly learn all skills … and the intended effect of such mitigation measures, if any.
However, the adaptation is a great occasion to make these methods really fit with the game concept and a crucial part in the opinion of the player regardless of shear efficiency (for example, I'd be more forgiving of this issue for TES if the trainers did not feel so "generic"). In my opinion, this is probably the hardest part, yet the one with the biggest payoff. For me, it generally requires lengthy sessions of research and brainstorming (then a few of the ideas stand-out and can be fleshed out). A good recent example of the adaptation of a game mechanic, in this instance "Respec", would be Kingdom of Amalur where they integrated this feature along with other key features of the game and it's main narrative.
Issue 3 - Balancing XP between skills
It is very doubtful any system that involve a "grow as you play" model will ever use a factor of 1 for all activities in regard to the XP increase. Not all things are equally easy to do or learn. For balancing purposes (and not realism), you will have to consider notably the frequency of use during a set amount of time (e.g. hitting with a weapon vs fishing). Since you already have a system in place, you have a good starting point for an eventual transition into another system.
Putting aside this inherent difficulty with balancing various activities between themselves, there is an additional difficulty that arises from balancing skills (or role) between themselves. In theory, each major/principal roles should take about the same effort and time to master, if not for realism, for fun and balance. There is a sort of trap here where you might be tempted to mitigate issue #2 by making some skills easier to learn compared to others… this can cheapen roles heavily dependant on that skill.
On that second point, the situation differ greatly with PZ where the two main group of skills (combat per weapon type, stealth) somewhat all raise at the same time within each group. As for the other skills, most of them are standalone or fairly close. This probably helps balancing skills between themselves.
The list could go on
There are a lot of other issues and considerations with "grow as you play" systems. I admit that those are tempting and cool, but they are not simple.
My personal suggestion, since you seem to be leaning toward "grow as you play" in your comment
This will be brief and dry, so it may seem "generic". I believe however that with some creativity it could be fleshed out.
A sort of hybrid system with the inclusion (and revamp) of vital stats (now : strength and stamina ; eventual : more could be added).
All your skill usage generate their own XP that make them "naturally" increase. All that XP also pool either in a common pool, or better yet in stats related pool.
The pool(s) can then be used to increase the player stats OR purchase training that would require time to complete. It may or may not require a skill book first.
If you choose multiple stats pools option, I suggest that going cross-stats might be possible but more costly if you want to reinforce role exclusivity, even though it's illogical if you use the training part.
Edit : Told you I'd go back to the multiple XP values
Edit 2 : Forgot to explain why I said your actual skill system is fairly OK. More on that at a later date and time.