I'm using the line-block construct extensively fro writing pseudocode. It usually works great, but boy does the rendered output look ugly. Here's an example of what one of my typical algorithms can end up looking in LaTeX:
\texttt{\ 1.}~\textbf{Inputs}:~variables~$A, b$\\\texttt{\ 2.}~\textbf{Output}:~$\phi$~~~~~~~~~~~~~~~~\texttt{//this\ is\ a\ comment}\texttt{\ 3.}~Choose~an~initial~guess~$\phi$~to~the~solution\\\texttt{\ 4.}~~\textbf{repeat}~until~convergence\\\texttt{\ 5.}~~~~\textbf{for}~$i$~\textbf{from}~1~\textbf{until}~$n$~\textbf{do}\\\texttt{\ 6.}~~~~~~~~$\sigma \leftarrow 0$\\\texttt{\ 7.}~~~~~~~~\textbf{for}~$j$~\textbf{from}~1~\textbf{until}~$n$~\textbf{do}\\\texttt{\ 8.}~~~~~~~~~~~~\textbf{if}~$j \ne i$~\textbf{then}\\\texttt{\ 9.}~~~~~~~~~~~~~~~$\sigma \leftarrow \sigma + a_{ij} \phi_j$\\\texttt{10.}~~~~~~~~~~~~\textbf{end~if}\\\texttt{11.}~~~~~~~~\textbf{end}~($j$-loop)\\\texttt{12.}~~~~~~~~$\phi_i \leftarrow \frac 1 {a_{ii}} (b_i - \sigma)$\\\texttt{13.}~~~~\textbf{end}~($i$-loop)\\\texttt{14.}~~~~check~if~convergence~is~reached\\\texttt{15.}~\textbf{end}~(repeat)
It's serviceable, but really hard to play with for a human. The HTML writer doesn't far much better. I realize that line-block are a special case where the usage of LineBreak can easily get out of control.
I wonder if it's better to add a newline after each LineBreak inline element. I'm really just targeting line-blocks, but as they don't have their own block-level type the only alternative is to change LineBreak behavior. The above example would then look like this:
\texttt{\ 1.}~\textbf{Inputs}:~variables~$A, b$\\
\texttt{\ 2.}~\textbf{Output}:~$\phi$~~~~~~~~~~~~~~~~\texttt{//this\ is\ a\ comment}
\texttt{\ 3.}~Choose~an~initial~guess~$\phi$~to~the~solution\\
\texttt{\ 4.}~~\textbf{repeat}~until~convergence\\
\texttt{\ 5.}~~~~\textbf{for}~$i$~\textbf{from}~1~\textbf{until}~$n$~\textbf{do}\\
\texttt{\ 6.}~~~~~~~~$\sigma \leftarrow 0$\\
\texttt{\ 7.}~~~~~~~~\textbf{for}~$j$~\textbf{from}~1~\textbf{until}~$n$~\textbf{do}\\
\texttt{\ 8.}~~~~~~~~~~~~\textbf{if}~$j \ne i$~\textbf{then}\\
\texttt{\ 9.}~~~~~~~~~~~~~~~$\sigma \leftarrow \sigma + a_{ij} \phi_j$\\
\texttt{10.}~~~~~~~~~~~~\textbf{end~if}\\
\texttt{11.}~~~~~~~~\textbf{end}~($j$-loop)\\
\texttt{12.}~~~~~~~~$\phi_i \leftarrow \frac 1 {a_{ii}} (b_i - \sigma)$\\
\texttt{13.}~~~~\textbf{end}~($i$-loop)\\
\texttt{14.}~~~~check~if~convergence~is~reached\\\texttt{15.}~\textbf{end}~(repeat)
I can't really think of any negative ramification off the top of my head. I'd appreciate any help thinking about whether this would break anything.