When trying to extract useful bits from HTML, a good general strategy is to use HTML::Parser to decompose HTML into its constituent elements and extract the parts you want with event handlers.

-Mark

Regarding the original question, one useful tool you may look into is the lynx used in conjunction with its -dump option. Combined with Perl regexps, there's not much you can't do with the actual information and links on the page.

     Hi kvale, you said that a good general strategy is to use HTML::Parser to decompose HTML into its constituent elements and extract the parts you want with event handlers..
     Even though this seems like a good way to handle HTML and retrieving data, I'm not convinced it's quite sufficient or efficient at all : I've been wanting to extract useful information from a webpage. What I infer by useful information is actually when you're on a newspaper website reading an article, to be able to retrieve the article only. In order to do that you need to find the beginning and the ending of the article's body. However, within the article itself there can be several HTML tags. I'm afraid your method would simply split the article apart turning it into nonsense.
     I haven't found any better way than to have a look at the HTML code itself and finding out whether special tags are used. Newspaper webmasters may sometimes use hidden HTML tags (<!-- article start-->) but then I need to come up with templates depending on which newspaper's website I'm currently analyzing.
     Have you any other idea ? I'd greatly appreciate your comments on this.