PROPOSAL: Literate Haskell in markdown noation

[Austin Seipp]

This argument is a little tenuous in my opinion. Markdown is very
ill-specified formally speaking with many implementations (including
our beloved Pandoc) featuring various extensions. It is not really a
'standard' in the typical sense. Pandoc actually only recently
supported 'GitHub code delimiters' - as in the above patch - only as
of a year ago or so, I believe.

In fact I believe Gruber's Markdown does not even specify a way to
delimit code blocks syntactically like this. So any way you slice it,
it's pretty "not standard." As for the MD vs TeX argument, well, it's
a pretty different tool entirely.

All that said, I'd support this patch. But mostly because it doesn't
actually introduce Markdown in any way, as much as it introduces a
relatively-simple new literate-style delimiter, in essence. (There are
arguably other forms we could support too, like the simple ``` form,
but that's where the ill-specification begins to rear its head...)

Finally, before a merge, it'd be nice to get this mentioned in the
users guide, and the release notes of course (I've been trying to keep
them a little more up to date.)

On Mon, Jun 24, 2013 at 3:46 PM, David Luposchainsky
<dluposchainsky at googlemail.com> wrote:
> I'm all for this. Over the years, many markup languages have come up,
> and juggling between Mediawiki, Haddoc, HTML, Latex and what not is
> really unnecessary. Markdown has reached a certain de-facto standard for
> beefed up text files. In the context of Haskell, it's easy to type (bird
> tacks are pretty awful to write multiple lines fluently with if you ask
> me), and it doesn't look as awkward than literate Tex Haskell, being
> readable in text and rendered formats.
>
> David
>
> _______________________________________________
> ghc-devs mailing list
> ghc-devs at haskell.org
> http://www.haskell.org/mailman/listinfo/ghc-devs



-- 
Regards,
Austin - PGP: 4096R/0x91384671