escape latex

Inlines require escaping with backticks

Displays require escaping with a div tag

content/notes/cpp.md

@@ -624,15 +624,15 @@ uint8_t var = -1;
 
 The value `-1` is encoded by first representing it as a positive number:
 
-$$ 0000\ 0001 $$
+<div>$$ 0000\ 0001 $$</div>
 
 The digits are then flipped, so that 1s become 0s and vice versa:
 
-$$ 1111\ 1110 $$
+<div>$$ 1111\ 1110 $$</div>
 
 Finally, the value is incremented by 1 to arrive at the Two's Complement representation of `-1`:
 
-$$ 1111\ 1111 $$
+<div>$$ 1111\ 1111 $$</div>
 
 When this value is assigned to an unsigned integer, the value is simply interpreted as if it were unsigned to begin with. Therefore, this value is interpreted as being `255`.
 
@@ -660,24 +660,26 @@ The modulus operation `%` simply calculates the remainder of the left expression
 
 The equation generally used to calculate the modulus is:
 
-$$ \text{mod}(a, n) = a - \lfloor a / n \rfloor * n $$
+<div>$$ \text{mod}(a, n) = a - \lfloor a / n \rfloor * n $$</div>
 
 The operation `-1 % 256` yields the result `255` with this implementation. This is the result yielded in languages such as Python and Ruby.
 
 C and C++ uses the same equation as the above, **but** the division operation has an additional restriction when used with negative operands:
 
-$$ \text{div}(-a, n) = \text{div}(a, -n) = -(a/n) $$
+<div>$$ \text{div}(-a, n) = \text{div}(a, -n) = -(a/n) $$</div>
 
 With these definitions, the division of `-1 / 256` in the above equation becomes `-(1 / 256)`. The result of `1 / 256` is zero due to truncation. The negation of this result is still zero, so the result of the modulus operation is simply `-1`, which is **very different** from the result of `256` yielded above without these restrictions.
 
 Given the above restriction on the division operation with negative operands, the definition of the modulus operation with negative operands can be simplified to:
 
+<div>
 $$
 \begin{align}
   \text{mod}(\phantom {-} a, -n) &= \phantom {-} \text{mod}(a, n) \\
   \text{mod}(-a, \phantom {-} n) &= -\text{mod}(a, n)
 \end{align}
 $$
+</div>
 
 # Exceptions
 
@@ -3729,7 +3731,7 @@ std::pair<int, int> bounds = std::minmax(std::rand() % v.size(),
                                          std::rand() % v.size());
 ```
 
-The `nth_element` function selects the $n^\text {th}$ element from the sorted order of the range, i.e. the $n^\text {th}$-order statistic. A comparison function can be specified.
+The `nth_element` function selects the `$n^\text {th}$` element from the sorted order of the range, i.e. the `$n^\text {th}$`-order statistic. A comparison function can be specified.
 
 This is like Quicksort's partitioning, so the range is modified so that the element pointed to by `nth` becomes the `nth` element of the sorted order of the range. All elements to the left of the iterator are less than or equal to that element and all elements to the right of the iterator are greater than that element.
 
@@ -3838,7 +3840,7 @@ v = {1, 1, 1, 1};
 
 The `partial_sum` function successively computes the sums of increasing sub-ranges of the input range and copies each sum into the output iterator. A custom sum function can be provided. Specifically, the result is such that:
 
-$$ \text {dest}[i] = \sum_0^i \text {src}[i] $$
+<div>$$ \text {dest}[i] = \sum_0^i \text {src}[i] $$</div>
 
 This can be useful for example to compute the [maximal sub-array](https://en.wikipedia.org/wiki/Maximum_subarray_problem).
 

content/notes/cryptography.md

@@ -6,7 +6,7 @@ date = 2016-05-04
 kind = "concept"
 +++
 
-Diffie-Hellman key exchange works by agreeing on two publicly shared values: a large prime number $q$ and a primitive root $g$. Alice and Bob each generate a _secret key_---a large random number---$x_a$ and $x_b$ respectively, and each raise the _primitive root_ to the power of the _secret key_, modulo the _large prime number_.
+Diffie-Hellman key exchange works by agreeing on two publicly shared values: a large prime number `$q$` and a primitive root `$g$`. Alice and Bob each generate a _secret key_---a large random number---`$x_a$` and `$x_b$` respectively, and each raise the _primitive root_ to the power of the _secret key_, modulo the _large prime number_.
 
 <div>
 $$
@@ -28,7 +28,7 @@ k_{ab} &= y_a^{x_b} \bmod q
 $$
 </div>
 
-Given a prime number $q$, a primitive root $g$ is a number such that every number from 1 up to $q - 1$ can be computed by raising the primitive root to some number $k$.
+Given a prime number `$q$`, a primitive root `$g$` is a number such that every number from 1 up to `$q - 1$` can be computed by raising the primitive root to some number `$k$`.
 
 <div>
 $$

content/notes/game-development.md

@@ -156,12 +156,13 @@ I learned linear algebra and related 3D math using math and Direct3D texts which
 
 ### Handedness
 
-The handedness of a system has an effect on the directions of the positive $z$-axis, positive rotation, and the cross-product. These can easily be determined by using the [right-hand rule](http://en.wikipedia.org/wiki/Right-hand_rule) or left-hand rule. It's important to realize that this has _nothing_ to do with row or column-major notation.
+The handedness of a system has an effect on the directions of the positive `$z$`-axis, positive rotation, and the cross-product. These can easily be determined by using the [right-hand rule](http://en.wikipedia.org/wiki/Right-hand_rule) or left-hand rule. It's important to realize that this has _nothing_ to do with row or column-major notation.
 
 ### Representation
 
-Row-major notation means that when indexing or denoting a matrix' size, the row is written first, and vice versa with column-major notation. Given a row-major matrix $R$, the column-major matrix $C$ is derived by deriving the transpose $R^{T}$:
+Row-major notation means that when indexing or denoting a matrix' size, the row is written first, and vice versa with column-major notation. Given a row-major matrix `$R$`, the column-major matrix `$C$` is derived by deriving the transpose `$R^{T}$`:
 
+<div>
 $$
 \begin{align}
 R &=
@@ -179,6 +180,7 @@ C = R^T &=
 \end{bmatrix}
 \end{align}
 $$
+</div>
 
 
 ### Storage
@@ -189,11 +191,11 @@ When it comes to matrix storage in memory, the key is that they are both stored
 
 When multiplying vectors with matrices, row-major notation requires pre-multiplication; the vector must appear to the left of the matrix. Column-major notation requires post-multiplication. This has the effect that the row-major, pre-multiplication transformation:
 
-$$ vMVP $$
+<div>$$ vMVP $$</div>
 
 Must be expressed as follows in column-major notation:
 
-$$ M^T V^T P^T v^T $$
+<div>$$ M^T V^T P^T v^T $$</div>
 
 However, row-major and column-major matrices are already transposes of each other, and as a result they are [stored](#storage) the same way in memory. This means that no actual transposing has to take place!
 
@@ -207,7 +209,7 @@ The most common optimization is to treat the world as a composition of _chunks_,
 
 A chunk may be represented by a volumetric grid such as a 3D array, optimized into a flat slab of memory indexed with the formula:
 
-$$ \text {chunk}[x + (y * \text {size}_x) + (z * \text {size}_x * \text {size}_y)] $$
+<div>$$ \text {chunk}[x + (y * \text {size}_x) + (z * \text {size}_x * \text {size}_y)] $$</div>
 
 The problem with this is that a lot of space is wasted for empty cells. A 16x16x16 chunk size where every cell conveys its information within one byte of data yields 4096 bytes, or 4 kilobytes. If every single cell in the chunk is empty except for one, then 4095 bytes are being wasted. That's 99.9% of the space allocated for the chunk.
 

content/notes/haskell.md

@@ -142,9 +142,9 @@ x <$ y = pure x <* y
 
 # Parallelism
 
-Amdahl's law places an upper bound on potential speedup that may be available by adding more processors. The expected speedup $S$ can be described as a function of the number of processors $N$ and the percentage of runtime that can be parallelized $P$. The implications are that the potential speedup increasingly becomes negligible with the increase in processors, but more importantly that most programs have a theoretical maximum amount of parallelism.
+Amdahl's law places an upper bound on potential speedup that may be available by adding more processors. The expected speedup `$S$` can be described as a function of the number of processors `$N$` and the percentage of runtime that can be parallelized `$P$`. The implications are that the potential speedup increasingly becomes negligible with the increase in processors, but more importantly that most programs have a theoretical maximum amount of parallelism.
 
-$$S(N) = \frac {1} {(1 - P) + \frac P N}$$
+<div>$$S(N) = \frac {1} {(1 - P) + \frac P N}$$</div>
 
 <img src="/images/notes/haskell/amdahls-law.png" class="center">
 

content/notes/java.md

@@ -1570,7 +1570,7 @@ The `LinkedList` class extends `AbstractSequentialList` and implements `List`, `
 
 The `HashSet` class extends `AbstractSet` and implements `Set`, and provides hash table behavior. Two of the constructors accept a capacity argument (default 16), with one of them accepting a load capacity argument (default 0.75) known as _fill ratio_, which determines how full the hash set can be before it is grown, and as such it must be a value between 0.0 and 1.0. In other words, the hash set is grown when:
 
-$$ \text{# of elements} \gt \text{capacity} \cdot \text{fill ratio} $$
+<div>$$ \text{# of elements} \gt \text{capacity} \cdot \text{fill ratio} $$</div>
 
 ### LinkedHashSet
 
@@ -1875,7 +1875,7 @@ There are also specialized variants of the `Optional` class for primitives which
 
 ## Random
 
-The `Random` class is a pseudorandom number generator. A variety of different kinds of numbers can be extracted from `Random` via different methods such as `nextBoolean`, `nextBytes`, `nextInt`, and so on. The `nextBytes` method in particular takes an array and fills it with the randomly generated values. The `nextInt` method has an overload that accepts an upper bound so that numbers are generated within the range $[0, n)$.
+The `Random` class is a pseudorandom number generator. A variety of different kinds of numbers can be extracted from `Random` via different methods such as `nextBoolean`, `nextBytes`, `nextInt`, and so on. The `nextBytes` method in particular takes an array and fills it with the randomly generated values. The `nextInt` method has an overload that accepts an upper bound so that numbers are generated within the range `$[0, n)$`.
 
 The seed can be passed to one of the constructor overloads or reset after the fact with the `setSeed` method.