I was going to go into an all caps rage on why 3rd edition skills are no worse than any other, but time cools my temper if not others. Especially after some thought. I've concluded this for now.
Subjectivity is one of the most useful and painful parts of role playing games. When a person in a undefined setting tells you that there is a clock, you gain an image of what that means to you. Perhaps you see a sun dial, perhaps a grandfather clock, or even perhaps a digital or building sized clock. Even within these definitions, should someone say it is a sun dial there are still variations, is it a great stone block in the ground that supports it? Is the sun dial stone itself? Is it one with just posts or an ornate bronze one with carefully embossed numbers? This is the nature of subjectivity within describing a concrete object. The fortunate thing for us is that all of the above clocks tell time, furthermore, should you decide to say it's a sun dial, we are now informed how it does it's job as well.
The real problem is that in role playing games we are constantly barraged with subjective responses to abstracts, imaginary constructs and even new laws of physics or logic. Magic cannot be seen in the real world, aside from movie magic versions implanted within us or magicians, it is practically an unknown. More effort is required to define it for the players in a 'tangible' and definite way. More effort is required to make it predictable. For instance, if I were to tell you a mage just caused fire to appear and harm you, in DnD there are a plethora of possibilities for this cause, also you all probably came up with a variety of forms for this to occur. Burning hands with it's fire erupting from fingers outstretched, Fireball with a small bead of energy erupting into a circular wave of destruction, or perhaps flaming arrows? Here in each instance we hearken back to the things we know. Is it pure happenstance the spells that persist throughout the editions are the ones that practically spell out their appearance in their name? I doubt it.
And what about new physics? What do you see in a bag of holding? How does it appear? Like a fish eye lens as light is forced to bend into a strange space? Or like a window into a room? How does it appear when the objects 'shift' to the top for easy acquisition? What does a sphere that annihilates everything look like? A black hole? Is there a swirl of air around it and a great sucking wind pulling towards it as the very air is pulled in? What is your reasoning that there isn't? Or that a ten foot pole held by characters is the only thing annihilated when they poke it, but when they touch it through gloves the entire person is? These subjective instances can result in conflicts. The conflicts can result in a feeling of inconsistency, dissatisfaction or even arguments that effect those who had no such conflict.
But the most problematic subjective thing in my opinion is the subjective nature of the abstracts. Believe it or not the numbers and statistics of the game are perhaps the worst of all. Let us bring to the fore the greatest offender, base statistics. This is a pain in the ass. Even within themselves there is a real problem with setting an agreed interpretation, but when extending into their derivative statistics things get worse. At first you have to name the buggers, strength, vitality, stamina, vigor, constitution, toughness... Many of these get used together in the same system! What makes toughness different than stamina? Furthermore, how does strength not effect constitution or depend on it? Imagine for a moment a person with an outstanding strength and miniscule constitution. How do they maintain said strength or even get it in the first place? There are stories of body builders with such strength they break their own bones! (And probably lose their strength after and are horribly deformed...) How could a sickly person have that strength? Intelligence and Wisdom? Where does one end and the other begin? They are all abstracts, subjective and all around messy.
But the game itself is centered around these subjective abstracts, and for all their mess, they are what make the game playable, both statistically and creatively. You can interpret those differences to make a genius who is absent minded (although memory is constantly put under intelligence yet somehow low wisdom causes this.) Or a glass cannon of melee, dishing out tons of melee or thrown damage, but unable to take a hit. You can set up skill challenges for someone strong or fast or smart or just plain tough. But, the naming isn't all that makes them subjective, establishing the effects and power level is highly subjective. What is average? What can an average person do? Who is average? If all the people in the fantasy world are above average compared to earth humans at something, is the average for their world or ours? Once we've established a baseline stat, how much variance is there between increments?
This comes up a lot. "This stat is above average! *I* can lift that, why can't my character?" or, "But you're at an 8 in intelligence! You would NEVER know that!" It's subjective, and conflicts will arise time and time again. But when extrapolated into the derived skills we come across an even more interesting and problematic situation.
Let us continue on with the idea of Dungeon and Dragon's idea of third edition skills in mind. John Doe has a baseline stat of 10 int. This is established as an average person's ability. Orc McGee has a 6, this is established as a mentally disabled level of intelligence, halving it would put him on par with animals. Both John Doe and Orc McGee are presented with the problem of building a crude stone wall. Orc McGee is given a few tools, and John Doe is given nothing. Following the system's logic of bonuses and penalties, Orc McGee gets a penalty of -2 for being dumb as the rocks he's using, and because John Doe is without tools, he gets the EXACT SAME penalty. But not being given some tools (which perhaps he could improvise for piling rocks strategically) he is reduced to the skill level of a mentally disabled person. Hell, if you give them both tools and throw some cold weather at John Doe it does the same thing.
This sounds like I would say this is a bad system. Arguably I would say it isn't by it's very nature, but due to subjective preconceptions we've established, we have concluded that it is. For instance, by looking into the system further we can get some more interesting information. Being a d20 system means that if one assumes that there is a chance for success with or without bonuses or penalties, being mentally disabled only reduces the chance of success by TEN PERCENT. Hell, assuming that we don't use common sense, the town cow has a 30% chance to have common knowledge about literature with their hefty penalty of -4. Thus subjective and abstract numbers can create conflicts. Things like this are where 'rules lawyers' start really moving and can really get into conflict with, or take advantage of 'role players'.
If skills sound bad, we can further extend the problem into the GM's territory. A GM who is too uncertain to take a stand can get into real trouble and put themselves right in the frying pan. If the monster has an intelligence of 6 would it really do that? Or even, what is the percentage chance that a mentally disabled person should have to build a rock wall? The rules book gives numbers, but often the rules book has an agenda and bends the rules to a subjective direction good for them and not necessarily you. For instance, following the earlier DnD reference, it shows a strong preference for specialization. The rogue is the trap master, but you better max out your spot and search or you're in trouble! Trap DCs can extend pretty high, many of the starting level traps having DCs of 20 and above! If you go into the game with an average wisdom and no skill in spot or search, you only have a 5% chance of preventing some serious hurt! Furthermore, if you want to learn a skill, you only can take skill points at half rate if your class isn't oriented towards that skill! You want to analyze a spell? Most classes cant even guess what that wizard is going to cast, even if they watched them cast it ten times in a row in the stock rules. If you put one rank into the skill and have an average intelligence, you only have a 30% chance of succeeding on even the most feeble and basic of spells. In fact, if you read into the core books to get their intent, the idea is that every character gets their moment to shine. This means that every other character is shit out of luck if they want to try. The agenda is specialization. Know your role and do it WAY better than everyone else. This can cause problems when the GM is not aware of this agenda.
Some games, such as Savage Worlds, offer the opportunity to succeed at skills while being the least optimal build possible for doing so! I could be the weakest most pathetic person on the field and have a chance to kill a giant monster in one blow! It's highly improbable, but it's there. The system's agenda is more about telling a story. You take your character, give them some vices and strengths and thrust them into a world with some generic abilities. The intent being that you will follow the path you want your story to go, and it will give you a chance to have your sickly tremor ridden character get the clinching shot and save the day. Because it makes a great story. The system is also ridden with exploits and inconsistencies that don't interfere with telling a great story, but might hurt a gritty realistic game.
Sure! Each system has it's strength! You probably knew that right? But each system's strengths are based on it's subjectivity as well as the actual system. Dungeons and Dragons could make all the skills MUCH more accessible for everyone with just a few changes. Hell in fourth edition they already have! They granted a bonus of half your level to all skills and have you just get a static bonus to all your skills you specialize in. You could easily plug that right into third edition and still play the game. You could also reduce the DCs for all skills, or automatically give successes for 'common knowledge'. A creative, thoughtful or at the least flexible GM could push the subjective bonuses, base skills, DCs and everything very far in any direction they want, making 18s in stats 'average' or setting all the DCs lower. It all depends on if you want to make the game harder or easier on your players.
If anything, the selection of the system you favor, should probably be the limits of bending the subjective portions of the game, rather than the stock game itself. Of coarse, by altering it too far from the base game, you begin to challenge the subjective views of your players, and therein lies more conflict. Perhaps that's part of the draw in me building my own system, and investigating new ones. A clean slate of preconceptions, any number of possibilities, seeing the visions and agendas of different game makers.