One of the important reasons I have left D&D3.5 and gone back to Classic D&D (and thence onto Labyrinth Lord) is simplicity versus complexity. I don't need too much realism - just enough for the rules to carry the story-telling, not so much that the rules dominate the game more than the banter back and forth between the players and DM.
Hit points versus hit locations? Hit points are easier to keep track of and can be universally applied to any NPC or monster.
Single table for saving throws or different saves for each situation? I actually liked the D&D3.0 innovation of having Fortitude/Willpower/Reflexes saving throws - it gave a good idea of how the PC was trying to avoid a terrible fate, and I could see how it can be applied in situations not covered in the Classic D&D saving throw table. However, they then made it a whole lot more complicated by making the DM calculate the DC of each save (especially by enemy spellcasters) rather than just looking it up on a table.
Skills outside combat - Again I kind of liked the initial approach of D&D3.0 with its range of possible skills. But then it had to bring more maths into it than a busy DM already juggling NPC actions, map locations and metaplot requirements wants to deal with. Also the idea of skill points not exactly equalling skill ranks and then trying to budget skill points was all a bit of a hassle, especially for NPCs. Perhaps I should be taking a step back towards NWP, where you either got it or you ain't got it? Interestingly Labyrinth Lord does not address this area of gameplay, but the Rules Cyclopedia has a section on "General Skills" which I ought to look at again.
Monsters and encounters - creating new monsters seemed to involve calculating skill ranks, stats (including charisma!) and feats in scale with the monster's HD as if it were a PC with so many levels. This drained my will to create new cool monsters for 3.0. And then constructing encounters was another headache, with Encounter Levels being a mixture of exponential calculations and fudging - no wonder many DMs decided to keep it simple and just use single monsters with the right Challenge Rating. Calculating XP for these encounters, particularly with a party of mixed levels was also a real problem. It's interesting that D&D 4E seems to have tried to remedy this, with relatively simple stat blocks and fixed XP rewards that are added together - I welcome this idea, even though it's not enough to tempt me to play D&D4E.
Timekeeping in combat and actions. Dear me, don't get me started on the difference between Free actions, Immediate Actions and Swift Actions. If you do something in combat, it takes a round. End of.
If you think I am disparaging the whole of D&D3.0, it is not my intention to do so. One of the things it did was with the SRD (system rules document), laying down what the core rules were - what everyone could expect from a 3.0 game. Everything else was optional. I suppose this had been implicit rather than explicit in previous editions - there had been the 1E and 2E PHB and DMG, and either Monster Manual or Monstrous Compendium (that folder with the hole-punched sheets? It seemed like a nice idea until you tried transporting it around between sessions) that could be considered the core rules. Other stuff, whether published as splat books, campaign setting books, magazine articles or on the internet, the DM and players could decide what to use and what to leave, but with the PHB and DMG, the players kind of knew what the heart of the game would be about. As the DM has the final say, even the core rules are optional in a way, if the DM wants to modify the system. But the core rules still provide a starting point of what it is the DM wants to change.
Additional stuff beyond the core rules has usually been part of RPGs, especially D&D where the company (whether TSR or WotC) has the money to produce extra products and the fan-base to sell enough of them to make it worthwhile. From Greyhawk: Supplement I to the latest 4E or Pathfinder fan-produced web material, I have always liked the idea of new stuff, even if it doesn't always work out as well as I hope.
Although I could go on about how options can get out of control and overwhelm the game (particularly with players going over the top with character optimization with dubious, untested source material), this is not my point. If they are options, it is up to the gaming group to decide what they will use and what will not be used, though I admit it is tempting to try to use too much.
With this blog, I have set out what core rules I use as my starting point. Everything beyond that is optional, whether it be new or variant rules on alignment, new character classes, new monsters, new magic items or new spells. I believe this is an indicator of a good rules system - a simple core, but it allows lots of options to be added on as the DM and players see fit.
Part of me is actually reluctant to produce a whole lot of new spells, monsters and magic items on this blog. I feel that there are more than enough out there and that DMs and players are quite capable of finding and using them (and perhaps converting them from one edition to another). There were something like 1200 wizard spells in 2E if you go through the supplements and sourcebooks, and as for monsters - 2100 in 2E, and over 3000 in 3.0 if you also include D20/OGL products. Any of these can be used in other editions (such as Classic D&D/LL) with a bit of conversion. However, I reckon I can produce that sort of material when it suits the campaign setting of Kaelaross, rather than simply for its own sake.