Re: header syntax.

On Nov 23, 2012, at 9:34 AM, Dave Pawson <dave.pawson@gmail.com> wrote:

> On 23 November 2012 14:27, Ryan Freebern <rfreebern@unionstmedia.com> wrote:
> 
>>> That would be my preference, starting from simple, clearly defined
>>> syntax and semantic, valid with current implementations, adding
>>> complexity as we feel is justified.
>> 
>> "Valid with current implementations" is only one half; "encompasses expected usage" is the other.
> 
> I haven't added that as an objective Ryan. The wiki outlines the
> syntax I hope we can cover,
> could you be more specific with that phrase please, encompasses expected usage

I mean that a casual markdown user shouldn't find that a majority of their documents are invalidated by a parser that implements our spec. Again, this is vague, but I don't think it serves anyone's purposes to release a spec that drastically alters the markdown landscape.

I agree that additional profiles would alleviate this problem, but I think their development needs to be active alongside the core spec. Each time we decide to restrict the syntax in the core spec, we should explicitly add the dropped syntax options to a potential additional profile document. In this case, if we choose to use

#+ ...\n

as our header syntax, we should add the variant with arbitrary trailing hash marks and the setext variant to potential profile documents immediately. Then, if usage surveys show them to be uncommon, those profiles can be dropped.

Perhaps these documents can be part of the wiki for now and codified later.

Ryan