Can we make Markdown headers semantically correct?

Question

HTML has a six level header system for document organization. The idea being that the <h1> level should have the page name, with <h2> subsections underneath. If further subsections are needed there are <h3>, <h4>, <h5>, and <h6>. So you could have some thing like

<h1>Page Name</h1>
  <h2>Sub-section</h2>
    <h3>Sub-sub-section</h3>
      <h4>Sub-sub-sub-section</h4>
        <h5>Sub-sub-sub-sub-section</h5>
          <h6>Sub-sub-sub-sub-sub-section</h6>
    <h3>Sub-sub-section 2</h3>

This provides a nice hierarchical organization. Stack Exchange sort of supports headers, but it’s entirely manual. And for some reason, two of the Markdown constructs create <h1> and <h2> headers, even in answers. So if people use the Markdown content as follows

Section 1
=========
Sub-section 1
-------------
Sub-section 2
-------------

If in the question body, we'd get

<h1>Question?</h1>
   <h1>Section 1</h1>
     <h2>Sub-section 1</h2>
     <h2>Sub-section 2</h2>
   <h2>COUNT Answers</h2>

and if in an answer, we'd see

<h1>Question?</h1>
  <h2>COUNT Answers</h2>
    <h1>Section 1</h1>
      <h2>Sub-section 1</h2>
      <h2>Sub-section 2</h2>

Eek! Either of those is wrong.

Note: I'm skipping a lot of content for brevity's sake. Presumably most people aren't composing their questions or answers entirely of headers. I'm leaving out the paragraphs, divs, etc.

So rather than have === and --- make h1 and h2 headers, could they instead be h2 and h3 in question bodies and h3 and h4 in answers?

Similarly, can you do one of the following?

• Demote #, ##, ###, and #### to <h3>, <h4>, <h5>, and <h6> in answers?
• Or demote # and ## to <h3> in answers and leave the others as is?

Either should change questions as well, but # would become <h2> instead of <h3> in questions.

The first better preserves the existing hierarchy while the second better preserves the intentions of people who currently use <h3> headers to denote top level sections of answers. Either would tend to lead to a lot of Markdown content being slightly wrong, as one preserves semantics while the other preserves display.

Note: a side issue is that search engine optimization works best if there is only one h1 per page and it has the same content as the page <title>.

Examples of problems that arise in the current system:

• On Stack Overflow: Is using 'heading' Markdown okay in answers?
• On Code Review: Discouraging gigantic headers in answers?

Note that both those problems would be fixed by this proposal. No one would use an <h1> or <h2> tag via Markdown in an answer anymore.

Answer 1

This is a question about mechanics, from a mechanic to non-mechanics. Normal people will see it as splitting hairs.

This should be about changing the style, so even this header 1 is not so enormous.

That looks hideous. So, why did the OP ask in the first place? Why not put more focus on the font style? Right or wrong, there is an explanation for why this even came up...

In the HTML coding world where Markdown came from, we get constantly lectured about how we're "supposed" to honor the DOM (Document Object Model) and use it properly for accessibility.

Understand DOM’s power in the browser object hierarchy:

Once you have a handle on the DOM hierarchy, its relationships to scripting and browsers, and its OOP features, you can go beyond simply writing more efficient scripts: You can use the objects of the DOM hierarchy as building blocks for far more sophisticated Web apps than you’ve ever imagined writing. You can create complex structured documents and employ sophisticated navigation techniques that can empower a Web app to the level of the most intricate in-house real-time database apps. You can even assemble virtual documents for your client’s Web users’ convenience from multiple sources and keep everything straight via DOM.

See the kind of pressure we software folk deal with?

It admittedly does make a difference to other application developers, browsers, and search engines. For instance, if you have a # header 1 in your question, Google might take more note of it than if you simply use **bold**.

There's a bigger issue

See how huge that # header 1 text is? We're "supposed" to use that in our posts. But no one does because it's so ridiculously huge. It's bigger than the title for the question itself!

More than whether the HTML DOM honors the hierarchies "properly", we need to get rid of that enormously atrocious front page headline font typeface and get fonts and sizes that make the Markdown hierarchies more attractive to use.