Prior discussion aside, let's look at the three defenses many posters seem to think viable: attack roll, save, SR.
An attack roll is trivial much of the time at the levels we're talking about. Quickened True Strike has come online 8 levels ago and WILL be used in concert with a potentially encounter-ending spell. That also gets rid of miss chances, btw. We're looking at touch ACs upwards of 40 to even have a slight chance of not being hit.
Saving throws seem to be the biggie here. Getting your save DC up beyond a certain point is kind of hard. Adding a save to negate would therefore quite invalidate the spell if it only does damage (look at the kind of save-or-dies you could have been throwing around forever by now). A save for half damage... maybe. I'd still only memorize a pure damage spell in a 9th (!) level slot if the damage dealt would be more or less unmitigable.
Spell Resistance is kind of weird, in that some creatures will be totally vulnerable to the spell (so no balancing factor at all), while against others you need to make a roll. In a core-only game, spell resistance is actually a meaningful defense against a spell attack (that's one reason why Conjuration is such an OP school). Out of core, spell resistance is rather easy to handle (such as by casting Assay Resistance beforehand), unless it is of a ridiculous value.
All in all, I'd say these defenses are kind of whacky. Using them as a balancing factor can quickly make the spell either a) not balanced (further) at all, or b) unattractive to ever use.
Tacking on a save should only be done if the primary effect is rather a bit on the strong side (such as 10 points of damage/CL PLUS daze for 1d4 rounds PLUS extra damage/effects vs. Undead).
Tacking on a ranged touch attack only forces the use of Quickened True Strike, which is a bit of a resource, but no big deal. A melee touch further needs a means of delivering the spell at range (such as the Spectral Hand spell), but doesn't decrease the spell's actual power - it only tells the caster how many hoops he has to jump through, either making the spell unattractive to him/her OR not doing a damn thing to balance it.
Allowing SR to defeat the spell introduces a number of effects largely based on the campaign (lots of evil outsiders? Don't use. Lots of more mundanely powerful foes? You're good to go), and on the books used (e.g. Spell Compendium or no).
None of these will have any impact on a campaign in which the usual iconic core spells are used as-is (which is the majority of campaigns, I'd guess). Namely, if you make the spell unattractive: not only in that it deals single target damage, and not even enough damage to truly irritate half the foes you'll be facing at this level; but ALSO in allowing a save or making the caster jump through some hoops to confirm the outcome - then they'll just whip out trusty old Gate or Shapechange and go to town.
BTW, to all those who think these spells shouldn't be used as a benchmark: the only thing we really need to know for this discussion is whether the OP's campaign allows Gate, Shapechange, Time Stop etc. If it does, these are the benchmark, because people in his campaign WILL be using them.