What is the HTML compressor using? It’s missing quite a few opportunities.

Take this input HTML:

  <!doctype html>
  <html lang="en-au">
    <head>
      <meta charset="utf-8">
      <title>This is a title</title>
    </head>
    <body class="foo">
      <h1>Well.</h1>
      <p>I wonder…</p>
    </body>
  </html>

It produces this output:

  <!DOCTYPE html><html lang="en-au"><head><meta charset="utf-8"><title>This is a title</title></head><body class="foo"><h1>Well.</h1><p>I wonder…</p></body></html>

It converted “doctype” to uppercase (equivalent, but bad for compression). It didn’t strip <head> and </head> as superfluous. It didn’t remove unnecessary quotes around attribute values. It didn’t remove the unnecessary </p></body></html> closing tags.

Here’s what I say it should have emitted:

  <!doctype html><html lang=en-au><meta charset=utf-8><title>This is a title</title><body class=foo><h1>Well.</h1><p>I wonder…

30 characters shorter, and structurally equivalent.