Thresher's system sounds a lot like the one from Ultima (I haven't played any White Wolf games, so I don't know how that system works) - take a couple of words of power and combine them to make a spell effect. The system I came up with is similar - I took the seeds presented in the ELH, modified them, added several more to account for additional effects (creation of space, altering time, etc.), and expanded the modifiers to accomplish more without having to "eyeball" or guesstimate things. What I ended up with was a reasonably well-balanced, level-based epic spellcasting system. You can research the individual seeds and cast them as single spells, or you can combine them for greater effect. We've got a list of about 70 spells now, both from an existing campaign converted to 3.5, and stuff I came up with on my own, and I've used those to tweak the system for balance. The costs in both gp and XP are greatly reduced (but still costly), making it more worthwhile to create and cast epic spells - cost is 10,000 gp/level, research time is one month per level above 9th, and XP cost is 1/25th of gp. Thus, a 21st level spell would cost 210,000 gp, take 11 months to research, and cost 8,400 XP. After that, you can cast it as often as you're allowed, paying only the normal XP costs for the spell, if any.
The thing I like about the Epic Spellcasting system is that it allows for unusual casting types (ritual, cooperative, sacrifice, power component) and for unusual and customized effects.
We've got that covered as well. Our system uses things like henges, monoliths, and rituals to lower the ending spell level.
The thing I dislike about it is, well, execution. I'd like a system like ichabod's

) ) or one that allows for reasonable limitations on mitigating factors and for a progression that actually tracks caster level rather than Spellcraft check bonus.
Got that covered too - level-based. The only real limitation we placed on the end level was that no level can go below 10th (obviously); XP cost is like that for creating magic items - you can't lose a level, but you can save up XP to burn instead of increasing the level. There are other costs you can add to casting a spell, like ability damage, and there is no limit on the amount (but if you can't pay the cost, the spell fails).
And I think it would be better for an epic spell to be worked out ahead of time, not composed by the player on the spot. A DM really needs time to think about the proposed spell and tinker with it.
Definitely. Our DM allows us to come up with normal spells and use them on the spot (and they can be tweaked afterwards), but I'd say that epic spells have to be worked over in full beforehand because of the sheer power level you're dealing with.
Spells are about a lot more than doing damage. Epic Spells shouldn't just be higher damage attack spells than 1st-9th level ones. Bigger guns is not the answer.
Cool **** is the answer. That's one reason the Epic Spellcasting system is so damn neat. You can do stuff like create a whole new species of creature - you can't do that with nonepic spells.
Amen to that. We do have some damaging spells, but most of the stuff we have is the really cool stuff, like raising mountains, creating demiplanes, and enslaving demons.
Like I said in the other epic-spellcasting thread, the rules will be available for free download sooner or later; if anyone wants a copy now to look them over and offer comments and/or playtesting, let me know. I'm currently away from home, but I should be back around Christmastime, so I'll be able to e-mail them out.