Re: [Docutils-users] "bad" HTML output?

[David G. <go...@py...>]

David Abrahams wrote:
> I just ran HTMLTidy over the output of
> 
>    python html.py test.txt test.html
>
> in the tools directory. It had lots of complaints, many of which
> look serious, though it only flagged them as warnings.
...
| 67 warnings, 0 errors were found!

Warnings mean "I'm not sure, but I think something may be wrong here"
and it's left to the user to judge.  Many of these cases are not
really problems, just HTMLTidy being overly critical.

> In particular, I notice lots of characters which appear to be
> invalid (possibly nuls).
...
| Character codes 128 to 159 (U+0080 to U+009F) are not allowed in HTML

That's because test.html is encoded in UTF-8.  Looks like HTMLTidy
doesn't understand the ``<?xml version="1.0" encoding="utf-8" ?>``
processing instruction at the beginning of the file.  Try the "-utf8"
option.

| even if they were, they would likely be unprintable control
| characters. Tidy assumed you wanted to refer to a character with the
| same byte value in the Windows-1252 encoding and replaced that
| reference with the Unicode equivalent.

Dangerous assumption.

> I'm not an HTML expert, which is why I use Tidy. Are these
> worth doing something about?

Some are, some aren't.

| : Doctype given is "-//W3C//DTD XHTML 1.0 Transitional//EN"
| : Document content looks like XHTML 1.0 Strict

HTMLTidy doesn't understand the transitional DTD?  Seems odd.

| test.html:16:1: Warning: <meta> element not empty or not closed
| :17:1: Warning: <meta> element not empty or not closed

These are real errors, now corrected.

| :23:1: Warning: <table> lacks "summary" attribute
...
| The table summary attribute should be used to describe
| the table structure. It is very helpful for people using
| non-visual browsers. The scope and headers attributes for
| table cells are useful for specifying which headers apply
| to each table cell, enabling non-visual browsers to provide
| a meaningful context for each cell.

The "summary" attribute is not required by the HTML 4 spec, just
recommended.  While I sympathize with its aim, I don't know of any way
to automatically generate a summary of a table.  Eventually a formal
"table" directive may be written, and it could have a "summary"
option.

| :85:24: Warning: <a> Anchor "table-of-contents" already defined

This seems to be because both the "id" attribute of the container
element and the "name" attribute of "<a>" elements are set to the same
thing (as specified in Appendix C of the XHTML spec,
http://www.w3.org/TR/xhtml1).  HTML 4 and XHTML want elements to use
the "id" attribute, but Netscape 4 only works with the "name"
attribute on "<a>" tags.  Perhaps the id and name attributes ought to
be on the same element though...  HTML is a mishmash; can't win.
Unless there's a problem with a real browser (not just a tool like
tidy), I don't see the need to fix this.

Thanks for taking the time to run HTMLTidy and bring these to our
attention.

-- 
David Goodger  <go...@py...>  Open-source projects:
  - Python Docutils: http://docutils.sourceforge.net/
    (includes reStructuredText: http://docutils.sf.net/rst.html)
  - The Go Tools Project: http://gotools.sourceforge.net/