GitHub Flavored Markdown
github.github.com
github.github.com
This also vindicates the "OSS authors don't owe nobody nothing!" crowd, as GitHub steps up and fixes the problems in question without requiring any participation by the original author of Markdown.
The two months since Github forked Markdown suggests to me Markdown isn't at critical mass, ergo Atwood's stewardship criticism seems misplaced. If he argued it could become _something_ beyond a better humane markup than those in the know, then he has a more valid criticism, and picking a fight. Seems underhanded to me.
Best I know, Gruber didn't intend for it to become something. It has only because it is more humane than most, and yet, it is pales in adoption compared the dearth of Word toolbar knockoffs.
The problem is that Markdown is far past critical mass, and is well into denouement. Everybody has the same major issues with its design, and are dissolutioned by the same outright bugs in Gruber's crap implementation. It doesn't help that the language is designed to be implemented as a regex substitution shitpile, as Gruber doesn't know the first thing about grammars.
He's been sitting on a pile of bugfixes for years that he refuses to release under his imprimatur. He hasn't blessed or even mentioned any of the dozen superior reimplementations of his spec. A few months ago he mentioned that he likes the idea behind Github's behavior-modifying patch for RDiscount, but not for his own use.
He really needs to post a new version of Markdown.pl with a disclaimer about its poor quality, and explicitly bless several of the reimplementations for actual use.
Everybody? Denouement? Critical mass? Please cite.
In one day on news.yc, I've seen two Markdown critiques, which is two more than I've seen, ever. This is a new development to a Markdown user. Please note the bugfixes he sat on.
And why this needs to be argued anymore: The BSD license gives you freedom. DVCS is about politics. Do what you need to do. If they don't like what you've done, fuck 'em you've voted in code, theory and premise.
"PHP Markdown" started as a straight transliteration of Markdown.pl, and he lists 123 user-visible changes that he made. The few of those changes that add features were to deal with warts in Markdown's design that made it impossible to not fuck things up: http://michelf.com/projects/php-markdown/
The PHP Markdown guy called out Gruber over this same bullshit 3 years ago: http://michelf.com/weblog/2006/state-of-markdown/
The bugfixes Gruber's been sitting on were mentioned on his stagnant mailing list: http://six.pairlist.net/pipermail/markdown-discuss/
The BSD license does gives you freedom, but not authority or the name, and Gruber spent years being hostile towards forkers/reimplementors, and then spent years simply ignoring it. Besides, the only thing the license would give your freedom regarding is Gruber's code, and you really don't want that.
I do this because I don't know of any way in vim to set a 'soft word wrap' so that lines are soft-wrapped at a specific column instead of just at the max width of the buffer. (Creating a fake vertical split is not what I want.)
One thing that I feel like Markdown is sorely missing is the ability to link to anchors within the same document. It would be nice if each heading had its own anchor tag.
Their ruby version seems to much better!
Perhaps it should be renamed "developer markdown" - DMD - and then github/stackoverflow/... should all have a standard markdown fork?
Also having a standard test suite that tests for XSS and other holes would be nice :)
Unless of course you're using Javascript to parse user generated markdown instead of doing it server-side, which is just silly.
Markdown is a formatting language, not a method to protect against XSS. It belongs in a different tier of your code.
I love GFM. I've been using it on userscripts.org for months. But the JS version at http://github.github.com/github-flavored-markdown/preview.ht... has holes that their ruby version does not have - http://github.com/mojombo/github-flavored-markdown/issues/#i...