···16 \item I have corrected a mistake in \textbf{Example: definitional lenses for finite ordinals}, where I previously did not qualify that the loops on $\mathcal{F}(m)$ only form a proper set when $m \geq 2$.
17 \item I have added an example, \textbf{Definitional lenses over discrete reflexive graphs}, in order to illustrate how the coproduct of reflexive graphs arises from lenses. I have also moved the definition of discrete reflexive graphs earlier as requested.
18 \item I have corrected a typographical error in \textbf{Characterising identification induction with dependent lenses} (formerly \textbf{Characterising identification induction with midpoint lenses}) in which $\otimes_A$ was given the type $A\times A\to B$ rather than $A\times A\to A$.
019 \item I have significantly shortened the introduction to reflect reasonable assumptions about the audience, as requested.
20 \item I have added a mention to the introduction of Observational Type Theory as an alternative design as requested.
21 \item I have renamed the \textbf{bivariant midpoint lens} to \textbf{unbiased dependent lens}.
···16 \item I have corrected a mistake in \textbf{Example: definitional lenses for finite ordinals}, where I previously did not qualify that the loops on $\mathcal{F}(m)$ only form a proper set when $m \geq 2$.
17 \item I have added an example, \textbf{Definitional lenses over discrete reflexive graphs}, in order to illustrate how the coproduct of reflexive graphs arises from lenses. I have also moved the definition of discrete reflexive graphs earlier as requested.
18 \item I have corrected a typographical error in \textbf{Characterising identification induction with dependent lenses} (formerly \textbf{Characterising identification induction with midpoint lenses}) in which $\otimes_A$ was given the type $A\times A\to B$ rather than $A\times A\to A$.
19+ \item I have corrected a typographical error in \textbf{Binary product of path objects} in which $\gB$ was written as $\gA$.
20 \item I have significantly shortened the introduction to reflect reasonable assumptions about the audience, as requested.
21 \item I have added a mention to the introduction of Observational Type Theory as an alternative design as requested.
22 \item I have renamed the \textbf{bivariant midpoint lens} to \textbf{unbiased dependent lens}.
+1-1
nodes/po-closure-properties.tex
···9899100 \begin{corollary}[Binary product of path objects]\label[corollary]{lem:bin-prod-po}
101- If $\gA$ and $\gA$ are path objects, then so is $\gA\times \gB$.
102 \end{corollary}
103104 \begin{proof}
···9899100 \begin{corollary}[Binary product of path objects]\label[corollary]{lem:bin-prod-po}
101+ If $\gA$ and $\gB$ are path objects, then so is $\gA\times \gB$.
102 \end{corollary}
103104 \begin{proof}
-2
reviews/review2-excerpts.md
···29Page 19: Definition 48 could use a citation or proof sketch for getting
30between 1/2 to 3/4.
3132-Page 20: type in Corollary 55, an A should be a B
33-34Page 27: Since the paper is generally working in an informal type theory
35style, it seems worth a comment that Definitional lens and related
36definitions cannot be an internal definition because they require a
···29Page 19: Definition 48 could use a citation or proof sketch for getting
30between 1/2 to 3/4.
310032Page 27: Since the paper is generally working in an informal type theory
33style, it seems worth a comment that Definitional lens and related
34definitions cannot be an internal definition because they require a