-
Notifications
You must be signed in to change notification settings - Fork 528
HTML5 update
I think we need a path forward for the generated markup. A few things are beginning to be outdated by HTML5. The problem is that some websites depend on the current markup, making all this a breaking change for those sites. And it's hard to evaluate how many site such changes would break.
My current idea is to make a big leap to update everything to HTML5 in one shot, with a configuration option to revert to the old markup. The big leap would make it noticeable so that things aren't breaking sneakily one by one at each update. Making a list of everything in need for an update would be a good start. So here is a tentative list, feel free to improve:
Currently using the rev attribute for back-links. This attribute is no longer in HTML5.
Currently using the align attribute for cell alignment. This attribute is no longer in HTML5.
HTML5 brings back attributes like type to change the counter in an ordered list. This attribute would be helpful for legal documents. Currently the only solution is to wrap the list in another element and style it with CSS.
HTML 5 introduces new elements. While Markdown doesn't try to convey everything that is possible with HTML, some of the new things might benefit having a Markdown-specific syntax.
The HTML <figure> element allows us to regroup some content being used for illustrative purpose, often images. The <figcaption> element lets us put a caption on a figure.
Idea: use "!" to create a figure:
[ Figure 1: some image ]
! 
[ Figure 2: some text ]
! paragraph
this is it
blah bla
[ Figure 3: some code ]
! ~~~
<html>
~~~
[ Figure 4: some table ]
! [ Table 1: passwords for all my friends ]
user | password
---- | --------
mike | mypass
[ Figure 5: nested figures ]
! [ Figure X: another figure ]
! [ Figure Y: yet another ]
! this is getting ridiculous
Now, just make the leading "!" and indentation optional for an image and you get this:
[ Figure 1: A simple image ]
And preceding the figure with a caption could be optional:
! some figure with text and without a caption
And we might want to allow the caption to appear below the figure:
[ Figure 1: A simple image ]
The Markdown Extra Extended fork has experimental support for figure via an alternative fence using the equals sign.
===
=== [Figure 1: A simple image]
Including a method for attaching classes and possibly IDs to figures would also be helpful. A figure containing a blockquote might be handled drastically different from a figure containing an image or a video.
The implicit figure option in Pandoc may be of interest.
Not really new in HTML5, but if we add captions for figures, it should work the same way for table.
- There seem to be two or three things here. One is implementing some HTML5 stuff (e.g., figure), and a second is updating the deprecated stuff (e.g., table align). I guess a third thing is how to deal with implementing the first or the second, but the implementation issue of updating the deprecated need not require implementing the new (obviously this point is already made here). Priority might be on the stuff that is deprecated, which would make the task easier, as opposed to one fell swoop.
- Also it would be great to have the updates in a way so that whatever might change in terms of the markdown syntax itself a search-and-replace would help with the old stuff out there.