Maybe you have a glorified view of my program. Basically, there were 8 factorial (over 40,000) combinations of levels. One of those was 87654321. Now obviously we know that this cannot be done in the game, since you need to beat 3 before 4, and 4 before 5, etc. All my program does finds the position of the numbers, and if it doesn't follow the rule, it deletes the combination. That's how I was left with 672 combinations.
As for the program you're describing, I don't have it. It could be coded, yes, but it's probably more complicated than you think. First of all, there's the problem of gathering the data. Manually we have to find the times for completing the levels, and for walking place to place. Once we have all of the times, then we could write a program. The program, in essence, would work the same way our minds do to solve the problem. It'll see if you have the whistle, and then use that time rather than the walking time, etc etc. The trouble comes when you have multiple places to do the same thing. For example, you could burn the bush by level 2 for the meat on the way to level 2, or, on the way to level 8 from level 2. Obviously it's the latter, not only geometrically, but also because of how much money we'd have at the time.
If you're not visualizing it yet... There was only 8 factorial combinations of possible level orders. Imagine how many more combos there are for getting the money, the heart pieces, the swords, and the items. Yes, a program could be written that'll test EVERY combination and just report back with the shortest time, but we'd need to know the data of everything, and even then the program would still be a pain to write.
With logic we're able to eliminate an amazing amount of the combinations. For example, if you need to go to a store by Level X, you know that there is store A,B, and C near that level, and you could stop at the one that's most on-path to the next place you're going. It's unncessary to even consider store X,Y, and Z on the other side of the overworld because you already know that that long of a detour would waste time.
Now imagine that Level X has store A,B,C and so does Level Y, and Level Z. If going to store B from level X is faster than be or C, and going to store A is faster from level Y, and store C is faster from level Z, and you could go to the store at any of these times, which route is faster?
Level X --> store B --> next location
Level Y --> store A --> next location
Level Z --> Store C --> next location
Let me make up numbers now.
Level X --> store B --> next location (4 screens)
Level Y --> store A --> next location (2 screens)
Level Z --> Store C --> next location (1 screen)
Now we know the time to go to the store is after Level Z. The program would have to every store combination for every level. And in a game as large as Zelda, that would be a bitch and a half to code. Here's why. h is for heart
Level X --> store B --> h --> next (7 screens)
Level Y --> store A --> h --> next (9 screens)
Level Z --> Store C --> h --> next (8 screens)
Now all of a sudden the Level X route is faster.
I would be for writing the program is the game were larger than it is. If the same game were 100 levels and the overworld was 100 times the size, writing a program would be faster than thinking about it logically. However, you'd still have to come up with all of the rule sets and gather all of the data yourself. I don't think I'm a good enough programmer to do this one. I could barely visualize how to code it. It'll be loops inside of loops inside of loops inside of loops. Gah.
We're practically there, already, though. I'm pretty damned certain you have to go 341. There aren't so many combinations after that. The two routes posted before are probably the best two. And if the sword upgrades is fastest, Michael's route is probably the way to go. I think whether or not to get the swords is the last major problem to solve. After someone comes up with that answer, we're set, probably. I say probably because these routes were created by humans, and not by a program that tried every combination. There very well may be a factor or a route we missed, or didn't consider. I doubt it, though, but it is possible. For example, until recently I didn't even consider getting the power bracelet for warping. Then when I did, I found a route geometrically shorter than Michael's. At the time, though, I was damned certain Michael's was the shortest geometrically. You could imagine my surprise when I found my route, but I wouldn't be surprised if any even faster route is found. I'm doubtful it could happen, but it may.
I talk too much.
I'm glad though that we have more people looking at this and thinking about it though. Maybe you guys will see something we missed. Once this vid is created, it will be amazing though, particularly for us. Folks who don't know the game as well, and don't know how much planning went into this, won't find it as amazing as we would.