HTML 4 BDTD?

[John W. Kennedy]

I'm in the process of de-frame-ing a website with a couple thousand
pages of static HTML, and I've been building a tool that works pretty
well, based on javax.swing.text.html.parser technology, which I've never
used before. Large parts of the website are HTML 3.2, and everything's
just ducky. But there are a good many pages that are HTML 4.0, and my
program goes completely ca-ca on them, because I'm stuck with only the
built-in html32.bdtd file.

A) Is there any good reason that Sun didn't make up an html401.bdtd file
yonks ago?

B) Has anyone an html401.bdtd file to share?

C) Is there any other solution available? (No XML-based tool is going to
come close to handling this stuff -- it's all hand-written--not by me--
and it was painful enough doing various text-based global fixes to make
it parse properly as 3.2. -- lots of <b><i>blah</b></i> and that sort of
thing.)

 

[Daniel Pitts]

I'm in the process of de-frame-ing a website with a couple thousand
pages of static HTML, and I've been building a tool that works pretty
well, based on javax.swing.text.html.parser technology, which I've never
used before. Large parts of the website are HTML 3.2, and everything's
just ducky. But there are a good many pages that are HTML 4.0, and my
program goes completely ca-ca on them, because I'm stuck with only the
built-in html32.bdtd file.

A) Is there any good reason that Sun didn't make up an html401.bdtd file
yonks ago?

B) Has anyone an html401.bdtd file to share?

C) Is there any other solution available? (No XML-based tool is going to
come close to handling this stuff -- it's all hand-written--not by me--
and it was painful enough doing various text-based global fixes to make
it parse properly as 3.2. -- lots of <b><i>blah</b></i> and that sort of
thing.)

--
John W. Kennedy
"The blind rulers of Logres
Nourished the land on a fallacy of rational virtue."
-- Charles Williams. "Taliessin through Logres: Prelude"

Check out JTidy (or just tidy). It'll clean up your HTML. It might
even be able to translate it to XHTML, and THEN you can use XML
parsing no problem

Standard java HTML parsing is very lacking (as you have discovered).
At the very worst, you may want to work with regex instead.

Oh, and see if Apache has anything (Maybe in Jakarta?), they tend to
have useful utilities of the most surprising type

Hope this helps,
Daniel.

 

[John W. Kennedy]

Check out JTidy (or just tidy). It'll clean up your HTML.

Yes, but I'm not trying to tidy it (though my current code does that as
a side effect, since I'm slurping each page into a tree and re-emitting
it in clean HTML4); I'm trying to do major surgery on the content of
every page, so that I can de-frame the whole website, which, although
elegant-looking to the user, has become a nightmare of frame-juggling
whenever I have to link from one page to another that is not a notional
child, parent, or sibling. The last thing I want to do is degrade the
existing HTML 4.0 pages (the majority of which are semantically
marked-up, thoroughly CSSed, and W3C verified) to HTML 3.2. I also want
a stable tool for future use, so that I can revise link menus in the
event of a new branch on the site's conceptual tree; otherwise, I'll
have to use SHTML for every single page.

It might
even be able to translate it to XHTML, and THEN you can use XML
parsing no problem

Maybe I'll have to do that, but I'm annoyed that I won't be able to use
real XHTML, but only XHTML-like HTML, thanks to Microsoft stabbing the
W3C in the back. (It's a public-oriented website, so I can't say "Use
Firefox", however much I'd like to.) I suppose I could make up the site
in XHTML and then XSLT it to an HTML4 equivalent.

Damn Microsoft! (And damn Apple for their cowardly acquiescence!)

 

[John W. Kennedy]

Check out JTidy (or just tidy).

On investigation, it appears to be able to be used as a library to read
HTML into a DOM. I'm more or less doing that now, so it should be
relatively straightforward to slot it in where I am using import
javax.swing.text.html, etc..

 

[Rogan Dawes]

On investigation, it appears to be able to be used as a library to read
HTML into a DOM. I'm more or less doing that now, so it should be
relatively straightforward to slot it in where I am using import
javax.swing.text.html, etc..

Also consider htmlparser (htmlparser.sourceforge.net)

Rogan

 

[John W. Kennedy]

Also consider htmlparser (htmlparser.sourceforge.net)

I looked at it, but liked the feel of JTidy better.

In practice, JTidy (as an in-program DOM-building tool, not as a
standalone application) has worked fine. I plugged it into my program,
replacing the javax.swing.text.html tools, in a few hours, and I can now
read HTML 4 and HTML 3.2 equally well. The end of the project to
de-frame the website and get all the pages 4.01-clean is now in sight.

I do wish the JavaDoc was a little more complete. In a few places, I had
to look at the source.