XML validation: how to check ALL validation problem for XHTML

Question

I have a lot of documents in HTML format (not very good) that I would like to convert in XML (XHTML). I know it is not so easy and I would use this strategy in a Java program:
1. Try to check the well-formness and validation with a XML parser (SAX or Xerces)
2. If not valid: try to individuate ALL the problems the file has (*and not only the first one that halts the processing process*)
3. Try to transform the HTML to a validable XHTML with some approach: regular expression or other methods
So the questions I do to you are the following:
1. What XML parser do you think is the best for this purposes? SAX or Xerces?
2. How can I understand what are all the validation problems in the file and not only the first one (If I remember well XML parsers halt the parsing process at the first error...)?
3. How can I transform the HTML to a valid XHTML? I have only to use RegEx or is there other tools to do this for XHTML and HTML problem?
Thanks
r
Edited by: robertobat on Feb 21, 2009 7:09 PM

Answer 1

>
1. Neither of them. (Disregarding the fact that SAX and Xerces aren't in the same category and don't cover all the possibilities.I would say SAX default implementation in JRE and SAX parser in Xerces.
2. I think you have "valid" and "well-formed" confused. And HTML isn't a dialect of XML so the idea of trying to use an XML parser to handle HTML isn't a good idea.I know very well what is the difference between valid and well-formed but I've used "validation" to represent all the conversion problem. But you are right. I'm convincing myself that using a XML parser as the first step is not a good idea.
3. Well, this is the real question, isn't it? Those other two were just a waste. Don't screw about with regex, for one thing it doesn't work well for hierarchical structures and for another you won't finish in a finite time. Just use an HTML parser which can produce a DOM, like TagSoup for example. Or run them through HTMLTidy. You could also submit them to one of the internet sites which will validate XHTML for you.I've seen Tidy and its capability to convert an HTML to a XHTML and I think it is better then TagSoup because I have to implement this mechanism in a production environment and I want to use only open source projects that have a very long story and that are strong. But I'll see TagSoup as you say.
I cannot use an Internet service to convert millions of private documents.