yes that was exactly the kind of magical argument I was refering to
Well, I probably sound like a broken record by now, but here is the basic of what I liked in different D&D editions. If you actually want to argue any of them, just say so:
OD&D: Although the many versions of the old D&D, from the White Box one to the Rules Cyclopedia, vary a lot in tone and content, their greatest strength is still the same I think. Basically, these books give you lots of little "set pieces", like spells, magic items, monsters, thief skills, dungeon building rules, keep rules, etc. They never, however, give you the "whole game". Instead, it is up to each game group to go creating rulings according to their approach to the exploration. Saying the system is good for not defining things, leaving the players to define them themselves, is stupid, of course. But the thing is that the set pieces help you flesh out the games and give you an idea of how to make your own rulings and house rules. Lots of these pieces suggest things, but never spell it out. Is your group interested in courtly intrigue and the politics of thieves' guilds? Then the GM will probably come up with good rules/tables/whatever to deal with these situations. Maybe something like Vornheim's nobleperson random table. You can see that too in the dungeon design, where the book is more concerned in making the GM understand what is in each room rather than explaining all possible consequences of the PC's action.
AD&D, 1st Edition: The first edition of AD&D suffers from a few poorly designed rules. The weapons vs AC rule, for example, could probably have been substituted by something a lot simpler. The organization of the book is pretty bad and it seems the designers are trying to lash out in every direction as to keep the players from making house rules themselves to cover possible, unexpected situations. Still, the "charm" here is that, in addition to the set pieces from before (to which this edition makes some additions), you also have a very good glimpse into how Gary and the other people in TSR approached their own rulings. The way he is always trying to create problems to the players while using verisimilitude as an excuse. The best way to play a long campaign of AD&D is, in my opinion, to wipe away most of the rules in the game so it is almost like OD&D again. Then, keep working on the rulings with the "Gygaxian way" in mind, making the world more detailed than what OD&D would have been but still keeping player conflict as the focus.
AD&D, 2nd Edition: This one has a lot of detractors, and I agree with them in a lot of things. Maybe this actually started mid-way through the 1st edition, but somewhere along the line, verisimilitude stopped being an excuse and started being a goal in itself. Now, this isn't necessarily a bad thing, but since the game had a very heavy luggage, it meant that the new priorities and the old ones would conflict, frequently making what was published for it not a lot of fun. And while some pretty awesome campaign settings date from this edition, the modules from this edition are also frequently very railroady. But the thing I like in this edition is when the meshing of the exploration of the setting and the challenging of the player characters do come together. For example, when the game pieces from the earlier editions are mixed with setting pieces of the new edition to create a setting that, while not completely serious or "realistic", is still very detailed and allows for the player's imaginations to be challenged. A concrete example of that can be seen in Planescape:Torment, where the various bits, like spells, weapons and proficiencies, were linked to the setting.