I don't get it. Say I already know that I'll need some formulae in my document. Then the use case here seems to be, essentially, that I can write
# Section Name
instead of
\section{Section Name}
and that I can write
*important stuff*
instead of
\emph{important stuff}.
I can see that the markdown version is a little nicer on the eyes and keyboard, but only by a small margin.
Now, if I write a document in latex I need to understand exactly
* latex.
Latex is a beast, but there are no serious alternatives to typeset formulae.
If I want to write markdown with some latex in it, I need to understand
* markdown
* latex
* how pandoc interleaves the two.
What if, as will inevitably be the case, something breaks? There will not even be close to as much documentation for the markdown+latex+pandoc stack as for the latex-only stack out there. Is there even a standard for markdown+latex, the pandoc way?
Then there's the issue of packages needed to compile. Tex suites are already huge pieces of software. Now I also need to have pandoc. Pandoc is written in haskell, and the haskell stack on arch is a complete clusterfuck. I will not have saved time if I need to understand how to fix pandoc if it breaks after an update. Granted, this is mostly an arch issue, but the point is that the more software you use, the more likely it is that something breaks.
The bottom line is that you replace an already complicated piece of software with the exact same piece of software and then something. I find it hard to justify it in this case where there is so little benefit.
Now, if I write a document in latex I need to understand exactly
* latex.
Latex is a beast, but there are no serious alternatives to typeset formulae.
If I want to write markdown with some latex in it, I need to understand
* markdown
* latex
* how pandoc interleaves the two.
What if, as will inevitably be the case, something breaks? There will not even be close to as much documentation for the markdown+latex+pandoc stack as for the latex-only stack out there. Is there even a standard for markdown+latex, the pandoc way?
Then there's the issue of packages needed to compile. Tex suites are already huge pieces of software. Now I also need to have pandoc. Pandoc is written in haskell, and the haskell stack on arch is a complete clusterfuck. I will not have saved time if I need to understand how to fix pandoc if it breaks after an update. Granted, this is mostly an arch issue, but the point is that the more software you use, the more likely it is that something breaks.
The bottom line is that you replace an already complicated piece of software with the exact same piece of software and then something. I find it hard to justify it in this case where there is so little benefit.