"Fixing html files"

[John Resler]

Hi all,
First I want to say I am fully aware of the huge scope of the problem
of parsing and correcting files of any sort. I have been using the jTidy
libraries (Dave Raggett W3C, I believe) to attempt to clean up the html
I use and convert it to xhtml if possible. Not to complain about Tidy,
it is the only application I'm aware of that does what it does... I am
just curious if there are any other applications/libraries that perform
the same function, more completely?

[David Carlisle]

John Resler <Jo*********@sbcglobal.net> writes:

Hi all,
First I want to say I am fully aware of the huge scope of the problem
of parsing and correcting files of any sort. I have been using the jTidy
libraries (Dave Raggett W3C, I believe) to attempt to clean up the html
I use and convert it to xhtml if possible. Not to complain about Tidy,
it is the only application I'm aware of that does what it does... I am
just curious if there are any other applications/libraries that perform
the same function, more completely?

Hard to quantify "more completely". tidy does a better job than most.
Alternative route might be for example John Cowan's tagsoup
http://mercury.ccil.org/~cowan/XML/tagsoup/
which will allow you to parse most html into an xml processing
pipeline. It doesn't do any cleaning up really, but once you have it as
xml you just hit it with enough xslt of your choice and it should all
come out looking lovely, er, in theory....

If you are feeling really brave there's my htmlparse xslt2 stylesheet
but this is decidedly unsupported.
http://www.dcarlisle.demon.co.uk/htmlparse.xsl

David

[Nick Kew]

John Resler wrote:

Hi all,
First I want to say I am fully aware of the huge scope of the
problem of parsing and correcting files of any sort. I have been using
the jTidy libraries (Dave Raggett W3C, I believe) to attempt to clean up
Dave Raggett wrote the original tidy, but it's been some years since
he was in charge of it.
the html I use and convert it to xhtml if possible. Not to complain
about Tidy, it is the only application I'm aware of that does what it
does... I am just curious if there are any other applications/libraries
that perform the same function, more completely?

libxml2 parses html, including tagsoup html, and gives you SAX or DOM
APIs on it. You can then serialise that to better HTML or XHTML.
It's a different approach to tidy, and shares the same fundamental
problem of having to guess blindly when presented with heavy-duty
gibberish.

A higher-level application based on libxml2 is AccessValet. Its
real purpose is (X)HTML accessibility analysis and reporting, but it
will also clean up (x)html. It takes a more brutal approach than
tidy: instead of attempting to substitute for crap, it strips it.
So if you take the default - which is strict output - it'll remove
everything that's deprecated in HTML4/XHTML1, and
<p align=center><font color=black>some text here<p>some more text
becomes
<p>some text here</p><p>some more text</p>

I wouldn't recommend it over tidy for that particular purpose, but it's
an option:-)

You can also fix markup on the fly when serving it. The state of the
art there is mod_publisher, at
http://apache.webthing.com/mod_publisher/
and is far better than any of the tidy-in-a-webserver options.

--
Nick Kew