About Esoth's WoW Hunter Tools
This site was created partly out of a desire to provide a computational tool that addresses a couple of concerns with existing tools. Broadly speaking, I have seen one type of problem with simulations like SimC and some web apps, and another type of problem with traditional spreadsheets. Let's ignore for a second the spreadsheet vs. sim debate (and questions about the computational power of a spreadsheet).
A common problem with the former is black box obfuscation: some calculations are being performed but you don't really know what they are. You insert an input into a black box (i.e. you can't see what's going on inside) and receive an output, but you don't know if it's "correct" or what kind of assumptions they are making. This makes it very difficult for users to identify bugs or run some tests using this modeling as a basis. Some apps literally hide everything. SimC is ostensibly not a black box in that the code is open source, but realistically most users aren't going to be able to identify problems or work around assumptions its making that they don't like without digging through a bunch of C code.
Spreadsheets on the other hand are generally pretty good at showing the math behind each calc. You can look at the formula for a given cell, and you might end up having to go through a bunch of them to figure everything out, but it is there. But these have a separate problem: distribution. SimC and spreadsheets both have the problem that even minor changes require an entirely new download (web apps do not have this problem). Additionally with spreadsheets, what if you have a bunch of settings you want to save? You've just required them to download a new copy, so those need to be redone unless you want to try to maintain profile files. Getting those to work across versions greatly increases the work needed to be done.
So my goal here was to create a web app that could "show its work" and, by nature of being a web app, always have the user using the newest version. I show my work partly by showing the details of how each stat and in turn how each spell is calculated. Theoretically someone who doesn't know programming could take a look and say "hey, this should/shouldn't use versatility here" without having to know the inner workings of the site. Similarly, the model view shows a shot sequence that the user can look at and see how certain changes actually change the model. I did use to make a spreadsheet awhile back, and there is one additional benefit of making a web app like this instead: using python is way more powerful and easy than trying to do ANYTHING in Visual Basic.
On Spreadsheets vs. Simulations
The general argument between these two approaches is generally that simulations are more accurate, but come at the expense of time. A single simulation is useless, but run it 10000 and you start to see the signal among the noise. The final result, although probabilistic, is something much more precise than a spreadsheet which works with averages. The spreadsheet will give more consistent results (not probabilistic) but they aren't as accurate. But it will give you this slightly rougher result much faster than a sim can give you anything useful.
This web app is not a sim and will never be a sim; for purposes of this argument is more along the lines of traditional spreadsheets. It calculates averages - if something has a 20% chance to proc, it will always use the average and assume it procs every 5th time. Basically, the goal of this app is to provide results that enable the user to make useful decisions without needing to get the more precise results of a sim.