I'm running AD&D right now, and in some ways, it's not so different. It's up to the DM to decide to what extent a fireball damages objects/ignites them. Sure, there's an item saving throw table, if the DM chooses to use it, for some or all objects caught in the area of effect. There
might be rules for room combustibility, AD&D having a lot of rules

, but most are really more like suggestions.
Okay, first of all, AD&D is what? 30 years old. There have been a number of changes (I would normally say improvements but I doubt we could agree on that term.) since then. Saying ....
A lack of consistency we have always had with us.
Isn't really true. It
was a lack of consistency (AD&D), then it became more consistent (3e) then it became less again (4e). A change that purposefully went from more information to less - as D'karr's prefers.
Let's assume the player has control for now...
If you're talking about 4e, they don't. The DM does.
What is "inconsistent" about a wizard having control over their own spell such that they can decide what objects, if any, are damaged by it?
Or, like I originally posted, why does a fireball have to behave like an incendiary grenade?
Because it isn't the wizard deciding these things, it is the DM. And it is the DM 's choice even when previously they ruled a different way and this time they change their mind.
Why, exactly, should a fireball spell function like an incendiary grenade or mortal shell? Other than it kinda used to in early editions?
Mortal shell? Not familiar with that term.
Why should it? Well should might be a strong term, but it
should because from a design standpoint they have to come up with a rule one way or another? Beyond that, as far as I know, there is no strong reason as to why it should behave like a grenade or any other type of explosion.
Why shouldn't the magician have control over what burns and/or ignites inside the area of effect. Or even if it does lethal or non-lethal damage?
Once again, from a design standpoint, it would have to pick if it deals lethal or non-lethal damage. Neither the player nor the DM should suddenly change their mind without redefining the actual rule involved. In specific however, the reason they have no control over what burns this time vs what burns next time has to do with the nature of the spell, the fact that those objects are in the path and that they should be consistent one way or another. Also because it isn't the players choice, it is the DMs's.
I'm not sure you actually are characterizing your tastes correctly. What you want is the game designers to tell you how the simulation works.
Agreed, let us continue.
The designers say that when Action A occurs, it will be resolved in this manner every single time.
Not that it will, but that it is supposed to - in favour of being consistent, yes. It would certainly suck if I kicked someone in the shin that this time they hurt their leg and next time they spontaneously ignite.
When you try to pick a lock, for example, in 3.5, you will use some sort of tool (taking penalties if the tools are jury rigged), it will take a certain amount of time, and you can retry as long as you want (thus taking 20, presuming you have time).
Right. Assuming you have plenty of time to do it, and the lock doesn't have a failsafe if you were to fail. Then sure I see no problem in taking 20 to open a lock.
And this is how Open Locks works every single time. Now, how about a puzzle lock, like a Chinese Puzzle Box? What skill do I roll to open that? Open Lock or Disable Device? I can see an argument for either way.
This is where you start to lose me. A puzzle box =/= a lock. But depending on circumstance either Open Lock or Disable Device may be appropriate.
In fact, if you found a chinese (they would not call it chinese) puzzle box in a (3e) DnD manual then it likely tells you what the DC is and which check should be used. 4e certainly wouldn't.
Now, in 4e, it's left up to the DM's discretion. You can open a lock in any manner that the DM accepts.
It was in 3e.
If you Fonzie Bump the lock, and the DM is happy, then groovy, you're just opened the lock. The rules expect you to make some effort towards doing just that. There's no compelling reason why you even have to use Thievery to open a lock, although that would be the baseline assumption. I could easily see a Wizard using Cantrip in conjunction with Arcana to open a lock as well.
Are you under the impression locks can't be destroyed in 3e?
What we are talking about is that the rule is use open lock to unlock the lock, DC 25 (for example). If you have Open Lock as a skill and roll (or take 20) then fine it is done. If you don't have Open Lock you can still destory it with any number of spells or weapons - probably.
What NEITHER rules support is the player "Fonzie Bump"-ing it. BOTH rules assume you have to actually try and unlock/disable the lock before it is opened.
Other DM's might not. And that's groovy. There's nothing saying that the DM must always say yes, although, again, that tends to be the baseline advice. But, there's certainly nothing stopping the DM from saying no. I could see a Martial Only campaign, doing a nice Sword and Sorcery 4e game, where that sort of thing just would not fly at all.
And here is the issue with the 4e ruling exactly. Some DM's may allow one effect, other DM's may allow another. Neither ruling is entirely supported by the game outside of "let the DM decide" on page 42. The problem arises when the DM allows the effect to do something one time and then changes it another time. Once again, this is perfectly allowed by the rules and is a major concern about realism and consistency. In 3e, when a fireball does (or doesn't) set paper aflame one time then everyone at the table can expect it do the same thing the next time it comes up. No DM fiat/judgement/ruling/interpretation/argument/opinion/best guess required. Similarly, if it is changed next time, then there is a reason and should be considered to work under those circumstances in the future.
In other words, because the mechanics are not lockstep tied to the flavor, you can apply different mechanics to an action without having to start jumping through all sorts of hoops. Yes, you could do the same thing in 3e, but, it was never encouraged to the degree that it is in 4e. If you want to do X, in 3e, you have to do Y. That's the formula. Departing from the formula is certainly not verbotten, of course, but, it's also not encouraged very much either.
This only matters, in either edition, if people at the game are 100% bound by the rules as written in the book. 4e doesn't solve that problem anymore than 3e did except it gave them less explanation on how effects logically occur. I feel it is poor design choice to give me less, especially on magical effects which so distort the usual understanding of cause and effect. Of course 4e opens it up to having more problems in this regard, as all effects are magical effects in 4e.
Also, about the emphasis mine; it is a formula for both - as many on both sides have pointed out.