Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
51 changes: 50 additions & 1 deletion BREAKING-CHANGES.md
Original file line number Diff line number Diff line change
Expand Up @@ -61,6 +61,22 @@ The calculus nodes and the binders — `derivative`, `integral`, `limit`, `sum`,
names the node and says to compile `.Evaled` or `Simplify()` instead. The LINQ compiler said the bound variable was not
among the arguments.

### A definite integral whose antiderivative splits into cases finds the breaks of each case

**Silent: a wrong value is corrected.** The antiderivative of `1/(t^2 + y^2)` is a case split on `y`, and its case
`y = 0`, `-1/t`, has a pole at 0. The breaks were looked for in the whole split, where the logarithm of the case
`y^2 < 0` cannot be listed, so none were looked for at all, and the case `y = 0` answered `-1/t` taken at the bounds:
`-2` over `[-1, 1]`, a negative value for a positive integrand whose integral diverges
([#2461](https://github.com/asc-community/AngouriMath/issues/2461)). Each case whose condition is on the parameters is now
taken across the range by itself, and the case `y = 0` diverges, as `integral(1/t^2, t, -1, 1)` does. The other cases
are as they were. With `Simplify`, the case `y = 0` of:

| Input | Was (2.7.0) | Now |
|---|---|---|
| `integral(1/(t^2+y^2), t, -1, 1)` | `(-2) provided (y = 0)` | `+oo provided (y = 0)` |
| `integral(1/(t^2+y^2), t, -oo, +oo)` | `0 provided (y = 0)` | `+oo provided (y = 0)` |
| `integral(1/(t^2+y^2), t, 1, 2)` | `(1/2) provided (y = 0)` | unchanged |

### `factorize`, `nextprime`, `divisors`, `gradient`, `divergence` and `φ` are refused by name

**Not silent: these names followed by a bracket now throw `UnrecognizedFunctionParseException` where they were read as
Expand Down Expand Up @@ -106,7 +122,40 @@ before. With `Simplify`:

The continuity read is on the real line with real parameters, as it is for finite bounds: for a non-real `y` with
`|Im y| > 1`, `arctan(t - y)` crosses arctan's cut and the integral of `1/(1 + (t - y)^2)` over the line is 0, not `pi`.
That was already the answer at a concrete such `y` -- `integral(1/(1+(t-3i)^2), t, -oo, +oo)` is `pi` in 2.7.0 and now.
At a concrete such `y` it is now 0: see the next entry.

### A definite integral whose antiderivative crosses a branch cut off the real line counts the jump

**Silent: a wrong value is corrected, and some are no longer answered.** The breaks of an antiderivative were looked for
on the real line, where `arctan` and the logarithm of a real argument have no cut to cross. A complex constant takes the
argument off the line, and there it can cross one where the integrand is continuous: `arctan(t - 3i)` meets the
imaginary axis at `-3i`, on arctan's cut, and jumps by `pi`, so the answer was off by `pi`
([#2460](https://github.com/asc-community/AngouriMath/issues/2460)). The crossings of arctan's cut and the logarithm's
are now placed for a polynomial argument with numeric coefficients, and for the logarithm of the cosine or the sine of a
linear one with numeric coefficients, which `tan` and `cot` integrate to, and each jump is counted; a bound on the cut
takes the value from inside the range. Where a crossing cannot be placed or its jump is not a constant, the integral is
left as written: a power whose base crosses the negative real axis, a bound that is not a number, a symbol beside the
complex constant, an argument of any other shape. With `Simplify`, against mpmath's quadrature:

| Input | Was (2.7.0) | Now | Value |
|---|---|---|---|
| `integral(1/(1+(t-3i)^2), t, -2, 2)` | `arctan(2 - 3i) - arctan(-2 - 3i)` = 2.8198 | `arctan(2 - 3i) - arctan(-2 - 3i) - pi` | -0.3218 |
| `integral(1/(1+(t-3i)^2), t, -oo, +oo)` | `pi` | `0` | 0 |
| `integral(1/(1+(t-2i)^2), t, -oo, +oo)` | `pi` | `0` | 0 |
| `integral(1/(1+(t-3i)^2), t, 0, +oo)` | `pi / 2 - arctan(-3i)` = 3.1416 + 0.3466i | `-1/2 * pi - arctan(-3i)` | 0.3466i |
| `integral(1/((1+i)*t-5-3i), t, 0, 5)` | `(1/2 - 1/2 * i) * ln(2i) + (-1/2 + 1/2 * i) * ln(-5 - 3i)` = 1.5510 + 2.6210i | `(1/2 - 1/2 * i) * ln(2i) + (-1/2 + 1/2 * i) * ln(-5 - 3i) + (-1 - i) * pi` | -1.5906 - 0.5206i |
| `integral(t/(1+(t-3i)^2), t, -2, 2)` | `3i * arctan(2 - 3i) + (-3i) * arctan(-2 - 3i) + (ln(-4 - 12i) - ln(-4 + 12i)) / 2` = 6.5670i | the same `+ (-2i) * pi` | 0.2838i |
| `integral(tan(t+i), t, 0, 4)` | `ln(cos(i)) - ln(cos(4 + i))` = 0.1376 - 2.4189i | `ln(cos(i)) - ln(cos(4 + i)) + 2i * pi` | 0.1376 + 3.8643i |
| `integral(tan((1+i)*t), t, 0, 4)` | `(-1/2 + 1/2 * i) * ln(cos(4 + 4i))` = -2.7951 + 0.5116i | `(-1/2 + 1/2 * i) * ln(cos(4 + 4i)) + (1 + i) * pi` | 0.3464 + 3.6532i |
| `integral(tan((2-i)*t+1/2), t, 0, 2)` | `(-2/5 - 1/5 * i) * ln(cos(9/2 - 2i)) + (2/5 + 1/5 * i) * ln(cos(1/2))` = -0.9264 + 0.4322i | the same `+ (2/5 - 4/5 * i) * pi` | 0.3302 - 2.0811i |
| `integral(sqrt((1+i)*t-5-3i), t, 0, 5)` | `4/3 * i + (-1/3 + 1/3 * i) * (-5 - 3i) ^ (3/2)` = 0.1675 - 5.3020i | left as written | 2.0531 - 3.4186i |
| `integral(1/(1+(t-i)^2), t, -1, 1)` | `arctan(1 - i) - arctan(-1 - i)` = 2.0344 | left as written | diverges: a pole at 0 |
| `integral(1/(1+(t-3i)^2), t, 0, s)` | `arctan((-6i + 2 * s) / 2) - arctan(-3i)` | left as written | |
| `integral(1/(1+(t-y-3i)^2), t, -2, 2)` | `arctan((4 - 6i + (-2) * y) / 2) - arctan((-4 - 6i + (-2) * y) / 2)` | left as written | |
| `integral(1/(1+(t-i/2)^2), t, -oo, +oo)` | `pi` | unchanged: the path stays inside the strip | `pi` |

A symbol is read as real here, as everywhere the breaks are read, so `integral(1/(1+(t-y)^2), t, -oo, +oo)` is still
`pi`, which is the value for `|Im y| < 1`.

### `dot` and `cross` of a name for a vector are nodes

Expand Down
Loading
Loading