The really nice thing about ReST is that it has provided generic syntax for extensions, one for inline text: :foo:`hello world` and for blocks:

.. extension:: hello world

In markdown, on the other hand, you have multiple, incompatible versions which have entirely different syntax because there is no generic extension mechanism.

ReST feels more well thought-out, generally.

That said, I've pretty much given up advocating it, because markdown seems to have won and has so much more tool support.

  +------------+------------+-----------+
  | Header 1   | Header 2   | Header 3  |
  +============+============+===========+
  | body row 1 | column 2   | column 3  |
  +------------+------------+-----------+
  | body row 2 | Cells may span columns.|
  +------------+------------+-----------+
  | body row 3 | Cells may  | - Cells   |
  +------------+ span rows. | - contain |
  | body row 4 |            | - blocks. |
  +------------+------------+-----------+

Pandoc is actually a really powerful tool for converting between different formats. As an example, I recently wrote a book using Markdown (Pandoc's version) and was easily able to export .html, .pdf and .epub from my markdown files. The addition of footnotes and built-in syntax highlighting (with optional line numbers, etc) was also very useful.

Pandoc can be as simple as `pandoc input.md output.pdf` but can also handle things like a Table of Contents, different highlighting styles, latex engines and fonts:

    pandoc --toc --variable version=0.0.1 -N --highlight-style=tango --latex-engine=xelatex --variable mainfont=Helvetica --variable monofont="Meslo LG L DZ" --chapters $(ls -d -1 `pwd`/_input/*.*) -o _output/book.pdf

I tend to set the more complicated command as a build file or alias and I've been considering using local markdown files and then using pandoc to convert them to .html for my WordPress-based blog.

The Markdown ecosystem reminds me of shell scripts in a bad way. There are a bunch of subtly and not so subtly different dialects and environments out there, and not many people have a strong awareness of these while writing, so if something seems okay "on my computer" or in Markdown's case "on my preview" then off it goes.

Am I the only one who feels like they were teleported back to the early 1990s upon seeing these text files with code intermingled everywhere?

I am not bashing markdown and friends, as I understand there is a use for these tools in some cases, but I am surprised they are so widely embraced and loved.

To me they just evoke the days of typing an essay on dad's 386 with Word Perfect 5.1 installed, and having to hit "reveal codes" to figure out what is going on. MS Word won against WP when they completely did away with these codes.[1]

Now it's 2014 and we're loving building tables by hand-crafting ASCII art.

Am I the only who thinks we can do better?

[1] http://www.theoligarch.com/microsoft_vs_apple_history.htm

Those ReST examples reminds me of the ASCII docs I'd write for my shareware software. Fairly typical for the time.

Not so different from the IEFT Document Conventions. http://www.rfc-editor.org/rfc/rfc3.txt

Or the RFC guidelines. https://www.rfc-editor.org/rfc-style-guide/rfc-style

I wrote a markdown renderer for my web dev stack. And I've been "cross compiling" to markdown, screen scrapping docs and persisting it to markdown.

Now I realize choosing markdown was rather arbitrary (personal preference, familiarity). Any document structure would suffice.

Nice comparison, thanks.

Markdown has always looked limited and incomplete to me.

But I can't decide between ReST(+Sphinx) and AsciiDoc(+?) - ReST seems to me like it was better thought out, but somehow my AsciiDoc documents turn out looking better, even though I like ReST more.

For me Markdown wins just because I'm used to dealing with it daily on GitHub.

Sidenote: Wish Github Flavored Markdown would adopt some nice things from Pandoc, like multiline tables.

Pandoc and ReST are extensive, flexible, and simply complete. They are great. I have never understood the fetish behind Markdown and the cult-like fan-boyish approach to it. It was not needed. It would have been rather good to standardised an existing markup (or a combination of them) than copy the existing one and pretty much release it with a different name.

Always wondered how Pandoc MD, ReST compare when my document type is slightly more complex, for example: a question and answer session?

I ended up entering notes from such sessions into handwritten XML, which I am reconsidering. What is a good format to store one's notes? I am transcribing from handwritten text.

i am in the process of releasing "zen markup language".

it's "lighter" than all the other light-markup languages.

it's more _powerful_ than the others, including asciidoc.

it's also far more agile, and much easier to understand.

and i won't allow it to be fragmented, like markdown is.

i've coded converters in javascript and other languages.

the javascript minimizes well for inclusion in web-pages.

i have cross-plat apps, and a web-app converter with a.p.i.

output formats include .html, .epub, .mobi, .pdf, and more.

if there's anything i haven't thought of, do please tell me.

you can reach me at my e-mail address given in my profile.

-bowerbird

Most the tools we have for WYSIWYG document creation lacks determinism and portability. For the latter, one could argue that there are open formats, but still, there should be a program that can interpret the format; whereas with plain text, such a need is void, as it is possible to view the document on any decent system. For the former, the argument might be that the modern word processors provide the facilities to deterministically lay out a document through the user interfaces they sport, but then because the formats they save are either are endemic to themselves or badly supported in other software* this feature is not of much use.

I can open a Markdown/ReST/Textile/... file with any text editor, including Notepad, Vim, Emacs etc., and also view it through more or less programs, or just cat them. I can pass them through head or tail; search them with common utilities and even edit them with some others. I am not bound to any programs in order to edit my program. If, on a computer I have to use, there is no Word, or Writer or Pages, I can still edit/read the document. I can read it online, via a browser. I can use programs that are decades old, and I also will be able to read the file decades later. When I send the file to someone else, I can be sure that they will be able to read it. Any usable operating system has a text editor bundled. This level of portability is just a dream for WYSIWYG editors. For these advantaged, I happily trade editing convenience off.

* Last summer, my cousins needed to use my computer for editing a docx document that was important for their undergraduate education. I was running Ubuntu OS at the time, so I told them to use the LibreOffice's word processor. The experience was bad; the document did not render properly, editing was problematic. This is the only case I can provide as an example to support my argument, as it has been multiple years since I used a word processor program.

The big gain is that is all just text and can be handled very well by the same version control system as the code is.

I can easily glance at what has happened to the docs just by looking at a diff, much like I do with all code my team produces already.

And being represented by all text doesn't mean this is what you give your customers. With a bit of work with pandoc etc. you can get a really slick and impressive end result.