I say keep the spellcasting in.
Now truth be told I totally missed the spellcasting aspect in my early attempts at running dragons of appreciable power. After all, no spell lists in the 3.0MM and when you run stuff on the fly....
Fast forward to now (more play experience, 3.5 = better Dragon stat blocks in the MM and the Draconomicon)) and I’ve not only caught on to what should have been obvious, but also learned a few things, such as:
*Dragon spellcasting is a means to throw in a handy spell after the fact that I might not have thought of before play (smells like DM cheating I’ll grant, but within limits this action serves to keep play vs. Dragons very challenging and fun).
*Spell selection and implementation is a means to add more flavor to each dragon, thus differentiating them in the player’s eyes and making each more unique.
*For an experienced Dungeon Master, a good Dragon encounter (including dragon spellcasting) can challenge every member of the party, which is as it should be. A dragon should be able to go head to head with the best fighter, sling and counter spells vs. the party wizard, keep the rogue on his toes and indenture the cleric to a full encounter’s worth of healing --all within one encounter. I don’t see this happening as often without the availability of spellcasting.
I personally feel Dragons are a corner case, inasmuch as game design vs. theme/sacred cows, etc... is concerned. That is, I just don’t think it’s a good idea to look only at game design and mechanics when thinking about how dragons and their abilities are detailed in game terms.
What’s more, I think Dave is off the mark so far as his vision of the archetypal Dragon is concerned; to me these creatures have always been as much about spells and magic as they are about immolating armies and rending multiple foes limb from limb.
As I understand it Dragons were deliberately designed to be more powerful than their CR would suggest. I say this idea should carry forward.
J. Grenemyer