Easy sharing.

Given the individual tasks' estimations (in hours), calculates the total estimation for the whole
project. It attempts to do the right thing, when summing up estimations, treating each individual
estimate as a continuous random variable (see *Math*), and applying a formula from Steve McConnel's
book "Software Estimation: Demystifying the Black Art".

Doing the math right though requires providing *a range* for an individual estimate instead of a
single number: you'll be asked for the best case, the most likely case, and the worst case hours.

By the way, the probability of a task taking exactly X hours to finish is exactly zero (see *Math*
again ;).

Also, a subjective *confidence factor* is provided for each task's estimation, in percent, within
the range (0, 100) -- excluding 0 and 100. Normally, you should be pretty confident in the *most* of
your tasks' estimations, specifying confidence factors above 90% for them. Only a smaller part of
tasks should have 90% and below. Anyway, that's a just fine tuning mechanism; you may well ignore
it, and simply assign, say, 95% confidence to all the tasks.

The number *E* you get from `total75PercentLikelyHours()`

means that there's a 75% probability that
your project will take not more than *E* hours.

Node.js:

```
$ npm install estimate-tasks
```

Web browser:

```
$ bower install estimate-tasks
```

Node version:

```
var et = require("estimate-tasks"),
total = et.total75PercentLikelyHours([{
bestCaseHours: 1,
mostLikelyCaseHours: 1.5,
worstCaseHours: 4,
confidencePercent: 99.9
}, {
bestCaseHours: 8,
mostLikelyCaseHours: 10,
worstCaseHours: 12,
confidencePercent: 66
}, {
bestCaseHours: 2,
mostLikelyCaseHours: 3,
worstCaseHours: 4,
confidencePercent: 95
}]);
```

In the browser you have to include the lib/calc.js script. Then:

```
var et = window["estimate-tasks"];
```

Test-driven with Mocha. Under Node, say:

```
~/estimate-tasks(master)$ npm test
```