Despite suggestions to the contrary, I don't "hate" skill systems or think they're anathema to old school game design. Rather, I don't see a lot of point in having a skill system in a class-based RPG, since they're either redundant or, worse yet, undermine the logic of classes. Consequently, when I play a class-based game, I generally assume that members of a given class can be expected to know about things related to that class. So, magic-users are knowledgeable about arcane lore and clerics are conversant in theology, etc. Specific character concepts, such as an illiterate, back woods wizard or a scholar-turned-fighting-man, might lead me to rearrange my assumptions a little, but, overall, I prefer to stick with them and view such specialized knowledge through the lens of character classes and run with it from there.
Now, over the course of my Dwimmermount campaign, the players have occasionally expressed an interest in their characters' learning something, such as a foreign language. Gaztea, Brother Candor's thief henchman, is in the process of learning Ancient Thulian, for example, and she's also experimenting with basic alchemy by virtue of the fact that we'd established she was a failed wizard's apprentice turned criminal (and has 17 Intelligence to boot). Since there's no formal way to handle the acquisition of such knowledge in the game, I've been winging it, expecting that time, money, and a tutor are what's needed for learning.
Then, just recently, I was re-reading my copy Empire of the Petal Throne and I saw a rule I'd forgotten about. Section 420 of the rulebook includes rules for "Original Skills," which are background skills not unlike the secondary skills of AD&D. At creation, a player rolls percentile dice to determine how many such skills his character starts with and from what categories. EPT has three categories: "plebeian," which covers ordinary arts and crafts, like baking and tailoring, "skilled," which covers more advanced arts and crafts, such as animal training and ship-building, and "noble," which covers very specialized knowledge requiring considerable study to acquire, such as alchemy or mathematics.
There are several things I like about EPT's "original skills" system. First, with very few marginal exceptions, the skills don't undermine the class system but rather complement it. Second, there's no universal mechanic associated with skill use. Possession of the skill brings with it no mechanical expectations; indeed, many of these skills have no means of resolution beyond referee fiat. And the skills that do have mechanics are tied closely to level, which I find quite agreeable. Finally, the system includes a means of acquiring new skills -- actually, it includes two. The first is based on level, as it's assumed characters will acquire new skills from various categories as they increase in level. The second is through the expenditure of time and money, with plebeian skills taking 2 months and 1000 gold pieces to learn, while noble skills take 6 months and 10,000 gold pieces.
EPT's system isn't without flaws and when/if I adopt it for use in my Dwimmermount campaign, I'll likely make some changes to it, but, taken as a whole, it's an approach to the question of specialized knowledge of which I approve. It's built to work in concert with the skill system without either weakening class archetypes or introducing mechanics uncongenial to my refereeing philosophy. I'm glad to was reminded of it.